diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index 7eede1ce..dbb73106 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -48,7 +48,7 @@ jobs:
           components: miri
 
       - name: Run miri
-        run: MIRIFLAGS=-Zmiri-ignore-leaks cargo miri test --features="alloc,defmt,mpmc_large,portable-atomic-critical-section,serde,ufmt,bytes"
+        run: MIRIFLAGS=-Zmiri-ignore-leaks cargo miri test --features="alloc,defmt,mpmc_large,portable-atomic-critical-section,serde,ufmt,bytes,zeroize"
 
   # Run cargo test
   test:
@@ -84,7 +84,7 @@ jobs:
           toolchain: stable
 
       - name: Run cargo test
-        run: cargo test --features="alloc,defmt,mpmc_large,portable-atomic-critical-section,serde,ufmt,bytes"
+        run: cargo test --features="alloc,defmt,mpmc_large,portable-atomic-critical-section,serde,ufmt,bytes,zeroize"
 
   # Run cargo fmt --check
   style:
@@ -176,6 +176,7 @@ jobs:
           cargo check --target="${target}" --features="portable-atomic-critical-section"
           cargo check --target="${target}" --features="serde"
           cargo check --target="${target}" --features="ufmt"
+          cargo check --target="${target}" --features="zeroize"
         env:
           target: ${{ matrix.target }}
 
diff --git a/CHANGELOG.md b/CHANGELOG.md
index ddb1d69b..14bb3511 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -13,6 +13,7 @@ and this project adheres to [Semantic Versioning](http://semver.org/).
 - Implement `defmt::Format` for `CapacityError`.
 - Implement `TryFrom` for `Deque` from array.
 - Switch from `serde` to `serde_core` for enabling faster compilations.
+- Implement `Zeroize` trait for all data structures with the `zeroize` feature to securely clear sensitive data from memory.
 
 ## [v0.9.1] - 2025-08-19
 
diff --git a/Cargo.toml b/Cargo.toml
index 76e2e0be..7390c86b 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -47,6 +47,9 @@ ufmt = ["dep:ufmt", "dep:ufmt-write"]
 # Implement `defmt::Format`.
 defmt = ["dep:defmt"]
 
+# Implement `zeroize::Zeroize` trait.
+zeroize = ["dep:zeroize"]
+
 # Enable larger MPMC sizes.
 mpmc_large = []
 
@@ -63,6 +66,7 @@ serde_core = { version = "1", optional = true, default-features = false }
 ufmt = { version = "0.2", optional = true }
 ufmt-write = { version = "0.1", optional = true }
 defmt = { version = "1.0.1", optional = true }
+zeroize = { version = "1.8", optional = true, default-features = false, features = ["derive"] }
 
 # for the pool module
 [target.'cfg(any(target_arch = "arm", target_pointer_width = "32", target_pointer_width = "64"))'.dependencies]
diff --git a/src/binary_heap.rs b/src/binary_heap.rs
index 22b882a1..452b84a1 100644
--- a/src/binary_heap.rs
+++ b/src/binary_heap.rs
@@ -17,6 +17,9 @@ use core::{
     ptr, slice,
 };
 
+#[cfg(feature = "zeroize")]
+use zeroize::Zeroize;
+
 use crate::vec::{OwnedVecStorage, Vec, VecInner, VecStorage, ViewVecStorage};
 
 /// Min-heap
@@ -55,6 +58,11 @@ impl private::Sealed for Min {}
 ///
 /// In most cases you should use [`BinaryHeap`] or [`BinaryHeapView`] directly. Only use this
 /// struct if you want to write code that's generic over both.
+#[cfg_attr(
+    feature = "zeroize",
+    derive(Zeroize),
+    zeroize(bound = "T: Zeroize, S: Zeroize")
+)]
 pub struct BinaryHeapInner<T, K, S: VecStorage<T> + ?Sized> {
     pub(crate) _kind: PhantomData<K>,
     pub(crate) data: VecInner<T, usize, S>,
@@ -882,6 +890,24 @@ mod tests {
         assert_eq!(heap.pop(), None);
     }
 
+    #[test]
+    #[cfg(feature = "zeroize")]
+    fn test_binary_heap_zeroize() {
+        use zeroize::Zeroize;
+
+        let mut heap = BinaryHeap::<u8, Max, 8>::new();
+        for i in 0..8 {
+            heap.push(i).unwrap();
+        }
+
+        assert_eq!(heap.len(), 8);
+        assert_eq!(heap.peek(), Some(&7));
+
+        // zeroized using Vec's implementation
+        heap.zeroize();
+        assert_eq!(heap.len(), 0);
+    }
+
     fn _test_variance<'a: 'b, 'b>(x: BinaryHeap<&'a (), Max, 42>) -> BinaryHeap<&'b (), Max, 42> {
         x
     }
diff --git a/src/c_string.rs b/src/c_string.rs
index 25df69b7..cf4ce301 100644
--- a/src/c_string.rs
+++ b/src/c_string.rs
@@ -10,6 +10,9 @@ use core::{
     ops::Deref,
 };
 
+#[cfg(feature = "zeroize")]
+use zeroize::Zeroize;
+
 /// A fixed capacity [`CString`](https://doc.rust-lang.org/std/ffi/struct.CString.html).
 ///
 /// It stores up to `N - 1` non-nul characters with a trailing nul terminator.
@@ -18,6 +21,20 @@ pub struct CString<const N: usize, LenT: LenType = usize> {
     inner: Vec<u8, N, LenT>,
 }
 
+#[cfg(feature = "zeroize")]
+impl<const N: usize, LenT: LenType> Zeroize for CString<N, LenT> {
+    fn zeroize(&mut self) {
+        self.inner.zeroize();
+
+        const {
+            assert!(N > 0);
+        }
+
+        // SAFETY: We just asserted that `N > 0`.
+        unsafe { self.inner.push_unchecked(b'\0') };
+    }
+}
+
 impl<const N: usize, LenT: LenType> CString<N, LenT> {
     /// Creates a new C-compatible string with a terminating nul byte.
     ///
@@ -483,6 +500,35 @@ mod tests {
         assert_eq!(Borrow::<CStr>::borrow(&string), c"foo");
     }
 
+    #[test]
+    #[cfg(feature = "zeroize")]
+    fn test_cstring_zeroize() {
+        use zeroize::Zeroize;
+
+        let mut c_string = CString::<32>::from_bytes_with_nul(b"sensitive_password\0").unwrap();
+
+        assert_eq!(c_string.to_str(), Ok("sensitive_password"));
+        assert!(!c_string.to_bytes().is_empty());
+        let original_length = c_string.to_bytes().len();
+        assert_eq!(original_length, 18);
+
+        let new_string = CString::<32>::from_bytes_with_nul(b"short\0").unwrap();
+        c_string = new_string;
+
+        assert_eq!(c_string.to_str(), Ok("short"));
+        assert_eq!(c_string.to_bytes().len(), 5);
+
+        // zeroized using Vec's implementation
+        c_string.zeroize();
+
+        assert_eq!(c_string.to_bytes().len(), 0);
+        assert_eq!(c_string.to_bytes_with_nul(), &[0]);
+
+        c_string.extend_from_bytes(b"new_data").unwrap();
+        assert_eq!(c_string.to_str(), Ok("new_data"));
+        assert_eq!(c_string.to_bytes().len(), 8);
+    }
+
     mod equality {
         use super::*;
 
diff --git a/src/deque.rs b/src/deque.rs
index f318859e..ec12bbf8 100644
--- a/src/deque.rs
+++ b/src/deque.rs
@@ -42,10 +42,14 @@ use core::marker::PhantomData;
 use core::mem::{ManuallyDrop, MaybeUninit};
 use core::{ptr, slice};
 
+#[cfg(feature = "zeroize")]
+use zeroize::Zeroize;
+
 /// Base struct for [`Deque`] and [`DequeView`], generic over the [`VecStorage`].
 ///
 /// In most cases you should use [`Deque`] or [`DequeView`] directly. Only use this
 /// struct if you want to write code that's generic over both.
+#[cfg_attr(feature = "zeroize", derive(Zeroize))]
 pub struct DequeInner<T, S: VecStorage<T> + ?Sized> {
     // This phantomdata is required because otherwise rustc thinks that `T` is not used
     phantom: PhantomData<T>,
@@ -1674,4 +1678,29 @@ mod tests {
 
         assert_eq!(Droppable::count(), 0);
     }
+
+    #[test]
+    #[cfg(feature = "zeroize")]
+    fn test_deque_zeroize() {
+        use zeroize::Zeroize;
+
+        let mut deque = Deque::<u8, 16>::new();
+
+        for i in 1..=8 {
+            deque.push_back(i).unwrap();
+        }
+        for i in 9..=16 {
+            deque.push_front(i).unwrap();
+        }
+
+        assert_eq!(deque.len(), 16);
+        assert_eq!(deque.front(), Some(&16));
+        assert_eq!(deque.back(), Some(&8));
+
+        // zeroized using Vec's implementation
+        deque.zeroize();
+
+        assert_eq!(deque.len(), 0);
+        assert!(deque.is_empty());
+    }
 }
diff --git a/src/history_buf.rs b/src/history_buf.rs
index b06e4648..67df4cd7 100644
--- a/src/history_buf.rs
+++ b/src/history_buf.rs
@@ -38,10 +38,14 @@ use core::ops::Deref;
 use core::ptr;
 use core::slice;
 
-mod storage {
-    use core::mem::MaybeUninit;
+#[cfg(feature = "zeroize")]
+use zeroize::Zeroize;
 
+mod storage {
     use super::{HistoryBufInner, HistoryBufView};
+    use core::mem::MaybeUninit;
+    #[cfg(feature = "zeroize")]
+    use zeroize::Zeroize;
 
     /// Trait defining how data for a container is stored.
     ///
@@ -83,6 +87,7 @@ mod storage {
     }
 
     // One sealed layer of indirection to hide the internal details (The MaybeUninit).
+    #[cfg_attr(feature = "zeroize", derive(Zeroize))]
     pub struct HistoryBufStorageInner<T: ?Sized> {
         pub(crate) buffer: T,
     }
@@ -148,6 +153,7 @@ use self::storage::HistoryBufSealedStorage;
 ///
 /// In most cases you should use [`HistoryBuf`] or [`HistoryBufView`] directly. Only use this
 /// struct if you want to write code that's generic over both.
+#[cfg_attr(feature = "zeroize", derive(Zeroize))]
 pub struct HistoryBufInner<T, S: HistoryBufStorage<T> + ?Sized> {
     // This phantomdata is required because otherwise rustc thinks that `T` is not used
     phantom: PhantomData<T>,
@@ -938,6 +944,39 @@ mod tests {
         assert_eq!(DROP_COUNT.load(Ordering::SeqCst), 3);
     }
 
+    #[test]
+    #[cfg(feature = "zeroize")]
+    fn test_history_buf_zeroize() {
+        use zeroize::Zeroize;
+
+        let mut buffer = HistoryBuf::<u8, 8>::new();
+        for i in 0..8 {
+            buffer.write(i);
+        }
+
+        assert_eq!(buffer.len(), 8);
+        assert_eq!(buffer.recent(), Some(&7));
+
+        // Clear to mark formerly used memory as unused, to make sure that it also gets zeroed
+        buffer.clear();
+
+        buffer.write(20);
+        assert_eq!(buffer.len(), 1);
+        assert_eq!(buffer.recent(), Some(&20));
+
+        buffer.zeroize();
+
+        assert_eq!(buffer.len(), 0);
+        assert!(buffer.is_empty());
+
+        // Check that all underlying memory actually got zeroized
+        unsafe {
+            for a in buffer.data.buffer {
+                assert_eq!(a.assume_init(), 0);
+            }
+        }
+    }
+
     fn _test_variance<'a: 'b, 'b>(x: HistoryBuf<&'a (), 42>) -> HistoryBuf<&'b (), 42> {
         x
     }
diff --git a/src/index_map.rs b/src/index_map.rs
index 34846240..e990c1e0 100644
--- a/src/index_map.rs
+++ b/src/index_map.rs
@@ -8,6 +8,9 @@ use core::{
     ops, slice,
 };
 
+#[cfg(feature = "zeroize")]
+use zeroize::Zeroize;
+
 use hash32::{BuildHasherDefault, FnvHasher};
 
 use crate::Vec;
@@ -66,6 +69,7 @@ use crate::Vec;
 pub type FnvIndexMap<K, V, const N: usize> = IndexMap<K, V, BuildHasherDefault<FnvHasher>, N>;
 
 #[derive(Clone, Copy, Eq, PartialEq)]
+#[cfg_attr(feature = "zeroize", derive(Zeroize))]
 struct HashValue(u16);
 
 impl HashValue {
@@ -80,6 +84,7 @@ impl HashValue {
 
 #[doc(hidden)]
 #[derive(Clone)]
+#[cfg_attr(feature = "zeroize", derive(Zeroize))]
 pub struct Bucket<K, V> {
     hash: HashValue,
     key: K,
@@ -88,6 +93,7 @@ pub struct Bucket<K, V> {
 
 #[doc(hidden)]
 #[derive(Clone, Copy, PartialEq)]
+#[cfg_attr(feature = "zeroize", derive(Zeroize))]
 pub struct Pos {
     // compact representation of `{ hash_value: u16, index: u16 }`
     // To get the most from `NonZero` we store the *value minus 1*. This way `None::Option<Pos>`
@@ -138,6 +144,11 @@ macro_rules! probe_loop {
     }
 }
 
+#[cfg_attr(
+    feature = "zeroize",
+    derive(Zeroize),
+    zeroize(bound = "K: Zeroize, V: Zeroize")
+)]
 struct CoreMap<K, V, const N: usize> {
     entries: Vec<Bucket<K, V>, N, usize>,
     indices: [Option<Pos>; N],
@@ -722,8 +733,14 @@ where
 ///     println!("{}: \"{}\"", book, review);
 /// }
 /// ```
+#[cfg_attr(
+    feature = "zeroize",
+    derive(Zeroize),
+    zeroize(bound = "K: Zeroize, V: Zeroize")
+)]
 pub struct IndexMap<K, V, S, const N: usize> {
     core: CoreMap<K, V, N>,
+    #[cfg_attr(feature = "zeroize", zeroize(skip))]
     build_hasher: S,
 }
 
@@ -1988,4 +2005,28 @@ mod tests {
         let map: FnvIndexMap<usize, f32, 4> = Default::default();
         assert_eq!(map, map);
     }
+
+    #[test]
+    #[cfg(feature = "zeroize")]
+    fn test_index_map_zeroize() {
+        use zeroize::Zeroize;
+
+        let mut map: FnvIndexMap<u8, u8, 8> = FnvIndexMap::new();
+        for i in 1..=8 {
+            map.insert(i, i * 10).unwrap();
+        }
+
+        assert_eq!(map.len(), 8);
+        assert!(!map.is_empty());
+
+        // zeroized using Vec's implementation
+        map.zeroize();
+
+        assert_eq!(map.len(), 0);
+        assert!(map.is_empty());
+
+        map.insert(1, 10).unwrap();
+        assert_eq!(map.len(), 1);
+        assert_eq!(map.get(&1), Some(&10));
+    }
 }
diff --git a/src/index_set.rs b/src/index_set.rs
index 53fd0ea3..803fa291 100644
--- a/src/index_set.rs
+++ b/src/index_set.rs
@@ -5,6 +5,9 @@ use core::{
     hash::{BuildHasher, Hash},
 };
 
+#[cfg(feature = "zeroize")]
+use zeroize::Zeroize;
+
 use hash32::{BuildHasherDefault, FnvHasher};
 
 use crate::index_map::{self, IndexMap};
@@ -85,6 +88,7 @@ pub type FnvIndexSet<T, const N: usize> = IndexSet<T, BuildHasherDefault<FnvHash
 ///     println!("{}", book);
 /// }
 /// ```
+#[cfg_attr(feature = "zeroize", derive(Zeroize), zeroize(bound = "T: Zeroize"))]
 pub struct IndexSet<T, S, const N: usize> {
     map: IndexMap<T, (), S, N>,
 }
@@ -696,4 +700,28 @@ mod tests {
 
     // Ensure a `IndexSet` containing `!Send` values stays `!Send` itself.
     assert_not_impl_any!(IndexSet<*const (), BuildHasherDefault<()>, 4>: Send);
+
+    #[test]
+    #[cfg(feature = "zeroize")]
+    fn test_index_set_zeroize() {
+        use zeroize::Zeroize;
+
+        let mut set: IndexSet<u8, BuildHasherDefault<hash32::FnvHasher>, 8> = IndexSet::new();
+        for i in 1..=8 {
+            set.insert(i).unwrap();
+        }
+
+        assert_eq!(set.len(), 8);
+        assert!(set.contains(&8));
+
+        // zeroized using index_map's implementation
+        set.zeroize();
+
+        assert_eq!(set.len(), 0);
+        assert!(set.is_empty());
+
+        set.insert(1).unwrap();
+        assert_eq!(set.len(), 1);
+        assert!(set.contains(&1));
+    }
 }
diff --git a/src/len_type.rs b/src/len_type.rs
index d42a659d..0eb6ff6a 100644
--- a/src/len_type.rs
+++ b/src/len_type.rs
@@ -3,6 +3,9 @@ use core::{
     ops::{Add, AddAssign, Sub, SubAssign},
 };
 
+#[cfg(feature = "zeroize")]
+use zeroize::Zeroize;
+
 pub trait Sealed:
     Send
     + Sync
@@ -75,6 +78,15 @@ macro_rules! impl_lentype {
 /// A sealed trait representing a valid type to use as a length for a container.
 ///
 /// This cannot be implemented in user code, and is restricted to `u8`, `u16`, `u32`, and `usize`.
+///
+/// When the `zeroize` feature is enabled, this trait requires the `Zeroize` trait.
+#[cfg(feature = "zeroize")]
+pub trait LenType: Sealed + Zeroize {}
+
+/// A sealed trait representing a valid type to use as a length for a container.
+///
+/// This cannot be implemented in user code, and is restricted to `u8`, `u16`, `u32`, and `usize`.
+#[cfg(not(feature = "zeroize"))]
 pub trait LenType: Sealed {}
 
 impl_lentype!(
diff --git a/src/lib.rs b/src/lib.rs
index 656999ee..fd506a92 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -106,6 +106,16 @@
 //! - [`mpmc::MpMcQueue`](mpmc): A lock-free multiple-producer, multiple-consumer queue.
 //! - [`spsc::Queue`](spsc): A lock-free single-producer, single-consumer queue.
 //!
+//! # Zeroize Support
+//!
+//! The `zeroize` feature enables secure memory wiping for the data structures via the [`zeroize`](https://crates.io/crates/zeroize)
+//! crate. Sensitive data can be properly erased from memory when no longer needed.
+//!
+//! When zeroizing a container, all underlying memory (including unused portion of the containers)
+//! is overwritten with zeros, length counters are reset, and the container is left in a valid but
+//! empty state that can be reused.
+//!
+//! Check the [documentation of the zeroize crate](https://docs.rs/zeroize/) for more information.
 //! # Minimum Supported Rust Version (MSRV)
 //!
 //! This crate does *not* have a Minimum Supported Rust Version (MSRV) and may make use of language
diff --git a/src/linear_map.rs b/src/linear_map.rs
index b7051737..c7f34278 100644
--- a/src/linear_map.rs
+++ b/src/linear_map.rs
@@ -4,6 +4,9 @@
 
 use core::{borrow::Borrow, fmt, mem, ops, slice};
 
+#[cfg(feature = "zeroize")]
+use zeroize::Zeroize;
+
 use crate::vec::{OwnedVecStorage, Vec, VecInner, ViewVecStorage};
 
 mod storage {
@@ -88,6 +91,11 @@ pub type OwnedStorage<K, V, const N: usize> = OwnedVecStorage<(K, V), N>;
 pub type ViewStorage<K, V> = ViewVecStorage<(K, V)>;
 
 /// Base struct for [`LinearMap`] and [`LinearMapView`]
+#[cfg_attr(
+    feature = "zeroize",
+    derive(Zeroize),
+    zeroize(bound = "S: Zeroize, K: Zeroize, V: Zeroize")
+)]
 pub struct LinearMapInner<K, V, S: LinearMapStorage<K, V> + ?Sized> {
     pub(crate) buffer: VecInner<(K, V), usize, S>,
 }
@@ -752,4 +760,24 @@ mod test {
         let map: LinearMap<usize, f32, 4> = Default::default();
         assert_eq!(map, map);
     }
+
+    #[test]
+    #[cfg(feature = "zeroize")]
+    fn test_linear_map_zeroize() {
+        use zeroize::Zeroize;
+
+        let mut map: LinearMap<u8, u8, 8> = LinearMap::new();
+        for i in 1..=8 {
+            map.insert(i, i * 10).unwrap();
+        }
+
+        assert_eq!(map.len(), 8);
+        assert_eq!(map.get(&5), Some(&50));
+
+        // zeroized using Vec's implementation
+        map.zeroize();
+
+        assert_eq!(map.len(), 0);
+        assert!(map.is_empty());
+    }
 }
diff --git a/src/sorted_linked_list.rs b/src/sorted_linked_list.rs
index d319fe0f..0b40b454 100644
--- a/src/sorted_linked_list.rs
+++ b/src/sorted_linked_list.rs
@@ -33,6 +33,9 @@ use core::mem::MaybeUninit;
 use core::ops::{Deref, DerefMut};
 use core::ptr;
 
+#[cfg(feature = "zeroize")]
+use zeroize::Zeroize;
+
 mod storage {
     use super::{LenType, Node, SortedLinkedListInner, SortedLinkedListView};
 
@@ -191,6 +194,7 @@ impl private::Sealed for Max {}
 impl private::Sealed for Min {}
 
 /// A node in the [`SortedLinkedList`].
+#[cfg_attr(feature = "zeroize", derive(Zeroize))]
 pub struct Node<T, Idx> {
     val: MaybeUninit<T>,
     next: Idx,
@@ -845,6 +849,26 @@ where
     }
 }
 
+#[cfg(feature = "zeroize")]
+impl<T, Idx, K, S> Zeroize for SortedLinkedListInner<T, Idx, K, S>
+where
+    T: Ord + Zeroize,
+    Idx: LenType + Zeroize,
+    K: Kind,
+    S: SortedLinkedListStorage<T, Idx> + ?Sized,
+{
+    fn zeroize(&mut self) {
+        while let Some(mut item) = self.pop() {
+            item.zeroize();
+        }
+
+        let buffer = self.list.borrow_mut();
+        for elem in buffer {
+            elem.zeroize();
+        }
+    }
+}
+
 #[cfg(test)]
 mod tests {
     use static_assertions::assert_not_impl_any;
@@ -974,6 +998,38 @@ mod tests {
         assert_eq!(ll.peek().unwrap(), &1001);
     }
 
+    #[test]
+    #[cfg(feature = "zeroize")]
+    fn test_sorted_linked_list_zeroize() {
+        use zeroize::Zeroize;
+
+        let mut list: SortedLinkedList<u8, Max, 8, u8> = SortedLinkedList::new_u8();
+        for i in 1..=8 {
+            list.push(i).unwrap();
+        }
+
+        assert_eq!(list.is_empty(), false);
+        assert!(list.is_full());
+        assert_eq!(list.peek(), Some(&8));
+
+        list.pop();
+        list.pop();
+        list.push(100).unwrap();
+
+        assert_eq!(list.peek(), Some(&100));
+
+        list.zeroize();
+
+        assert_eq!(list.peek(), None);
+        assert!(list.is_empty());
+
+        unsafe {
+            for node in &list.list.buffer {
+                assert_eq!(node.val.assume_init(), 0);
+            }
+        }
+    }
+
     fn _test_variance<'a: 'b, 'b>(
         x: SortedLinkedList<&'a (), Max, 42, u8>,
     ) -> SortedLinkedList<&'b (), Max, 42, u8> {
diff --git a/src/string/mod.rs b/src/string/mod.rs
index b7f07712..207d35c6 100644
--- a/src/string/mod.rs
+++ b/src/string/mod.rs
@@ -11,6 +11,9 @@ use core::{
     str::{self, Utf8Error},
 };
 
+#[cfg(feature = "zeroize")]
+use zeroize::Zeroize;
+
 use crate::CapacityError;
 use crate::{
     len_type::LenType,
@@ -142,6 +145,7 @@ pub type ViewStorage = ViewVecStorage<u8>;
 ///
 /// In most cases you should use [`String`] or [`StringView`] directly. Only use this
 /// struct if you want to write code that's generic over both.
+#[cfg_attr(feature = "zeroize", derive(Zeroize), zeroize(bound = "S: Zeroize"))]
 pub struct StringInner<LenT: LenType, S: StringStorage + ?Sized> {
     vec: VecInner<u8, LenT, S>,
 }
@@ -1462,4 +1466,26 @@ mod tests {
         let mut s: String<8> = String::try_from("a").unwrap();
         _ = s.insert_str(2, "a");
     }
+
+    #[test]
+    #[cfg(feature = "zeroize")]
+    fn test_string_zeroize() {
+        use zeroize::Zeroize;
+
+        let mut s: String<32> = String::try_from("sensitive_password").unwrap();
+
+        assert_eq!(s.as_str(), "sensitive_password");
+        assert!(!s.is_empty());
+        assert_eq!(s.len(), 18);
+
+        s.truncate(9);
+        assert_eq!(s.as_str(), "sensitive");
+        assert_eq!(s.len(), 9);
+
+        // zeroized using Vec's implementation
+        s.zeroize();
+
+        assert_eq!(s.len(), 0);
+        assert!(s.is_empty());
+    }
 }
diff --git a/src/vec/mod.rs b/src/vec/mod.rs
index 52a5f830..5e916dfb 100644
--- a/src/vec/mod.rs
+++ b/src/vec/mod.rs
@@ -11,6 +11,9 @@ use core::{
     slice,
 };
 
+#[cfg(feature = "zeroize")]
+use zeroize::Zeroize;
+
 use crate::len_type::{check_capacity_fits, LenType};
 use crate::CapacityError;
 
@@ -84,7 +87,11 @@ mod storage {
             Self: VecStorage<T>;
     }
 
+    #[cfg(feature = "zeroize")]
+    use zeroize::Zeroize;
+
     // One sealed layer of indirection to hide the internal details (The MaybeUninit).
+    #[cfg_attr(feature = "zeroize", derive(Zeroize))]
     pub struct VecStorageInner<T: ?Sized> {
         pub(crate) buffer: T,
     }
@@ -208,6 +215,7 @@ pub use drain::Drain;
 ///
 /// In most cases you should use [`Vec`] or [`VecView`] directly. Only use this
 /// struct if you want to write code that's generic over both.
+#[cfg_attr(feature = "zeroize", derive(Zeroize), zeroize(bound = "S: Zeroize"))]
 pub struct VecInner<T, LenT: LenType, S: VecStorage<T> + ?Sized> {
     phantom: PhantomData<T>,
     len: LenT,
@@ -2247,6 +2255,69 @@ mod tests {
             <alloc::vec::Vec<u8> as TryInto<Vec<u8, 1>>>::try_into(av.clone()).unwrap_err();
     }
 
+    #[test]
+    #[cfg(feature = "zeroize")]
+    fn test_vec_zeroize() {
+        use zeroize::Zeroize;
+
+        let mut v: Vec<u8, 8> = Vec::new();
+        for i in 0..8 {
+            v.push(i).unwrap();
+        }
+
+        for i in 0..8 {
+            assert_eq!(v[i], i as u8);
+        }
+
+        v.truncate(4);
+        assert_eq!(v.len(), 4);
+
+        for i in 0..4 {
+            assert_eq!(v[i], i as u8);
+        }
+
+        v.zeroize();
+
+        assert_eq!(v.len(), 0);
+
+        unsafe {
+            v.set_len(8);
+        }
+
+        for i in 0..8 {
+            assert_eq!(v[i], 0);
+        }
+    }
+
+    #[test]
+    #[cfg(feature = "zeroize")]
+    fn test_vecview_zeroize() {
+        use zeroize::Zeroize;
+
+        let mut v: Vec<u8, 8> = Vec::new();
+        for i in 0..8 {
+            v.push(i).unwrap();
+        }
+
+        let view = v.as_mut_view();
+
+        for i in 0..8 {
+            assert_eq!(view[i], i as u8);
+        }
+
+        view.zeroize();
+
+        assert_eq!(view.len(), 0);
+
+        unsafe {
+            view.set_len(8);
+        }
+
+        for i in 0..8 {
+            assert_eq!(view[i], 0);
+        }
+    }
+
     fn _test_variance<'a: 'b, 'b>(x: Vec<&'a (), 42>) -> Vec<&'b (), 42> {
         x
     }