diff --git a/fuzz/Cargo.toml b/fuzz/Cargo.toml
index 27040d7bc..f280d2807 100644
--- a/fuzz/Cargo.toml
+++ b/fuzz/Cargo.toml
@@ -57,6 +57,12 @@ path = "fuzz_targets/http_api.rs"
 test = false
 doc = false
 
+[[bin]]
+name = "pmem"
+path = "fuzz_targets/pmem.rs"
+test = false
+doc = false
+
 [[bin]]
 name = "qcow"
 path = "fuzz_targets/qcow.rs"
diff --git a/fuzz/fuzz_targets/pmem.rs b/fuzz/fuzz_targets/pmem.rs
new file mode 100644
index 000000000..79ad35b19
--- /dev/null
+++ b/fuzz/fuzz_targets/pmem.rs
@@ -0,0 +1,150 @@
+// Copyright © 2022 Intel Corporation
+//
+// SPDX-License-Identifier: Apache-2.0
+
+#![no_main]
+
+use libc::{MAP_NORESERVE, MAP_PRIVATE, PROT_READ, PROT_WRITE};
+use libfuzzer_sys::fuzz_target;
+use seccompiler::SeccompAction;
+use std::ffi;
+use std::fs::File;
+use std::io;
+use std::os::unix::io::{AsRawFd, FromRawFd, RawFd};
+use std::sync::Arc;
+use virtio_devices::{Pmem, UserspaceMapping, VirtioDevice, VirtioInterrupt, VirtioInterruptType};
+use virtio_queue::{Queue, QueueT};
+use vm_memory::guest_memory::FileOffset;
+use vm_memory::{bitmap::AtomicBitmap, Bytes, GuestAddress, GuestMemoryAtomic, MmapRegion};
+use vmm_sys_util::eventfd::{EventFd, EFD_NONBLOCK};
+
+type GuestMemoryMmap = vm_memory::GuestMemoryMmap<AtomicBitmap>;
+
+const QUEUE_DATA_SIZE: usize = 4;
+const MEM_SIZE: usize = 256 * 1024 * 1024;
+const PMEM_FILE_SIZE: usize = 128 * 1024 * 1024;
+// Max entries in the queue.
+const QUEUE_SIZE: u16 = 256;
+// Guest physical address for descriptor table.
+const DESC_TABLE_ADDR: u64 = 0;
+const DESC_TABLE_SIZE: u64 = 16_u64 * QUEUE_SIZE as u64;
+// Guest physical address for available ring
+const AVAIL_RING_ADDR: u64 = DESC_TABLE_ADDR + DESC_TABLE_SIZE;
+const AVAIL_RING_SIZE: u64 = 6_u64 + 2 * QUEUE_SIZE as u64;
+// Guest physical address for used ring (requires to 4-bytes aligned)
+const USED_RING_ADDR: u64 = (AVAIL_RING_ADDR + AVAIL_RING_SIZE + 3) & !3_u64;
+
+fuzz_target!(|bytes| {
+    if bytes.len() < QUEUE_DATA_SIZE || bytes.len() > (QUEUE_DATA_SIZE + MEM_SIZE) {
+        return;
+    }
+
+    let mut pmem = create_dummy_pmem();
+    let queue_data = &bytes[..QUEUE_DATA_SIZE];
+    let mem_bytes = &bytes[QUEUE_DATA_SIZE..];
+
+    // Setup the virt queue with the input bytes
+    let q = setup_virt_queue(queue_data.try_into().unwrap());
+
+    // Setup the guest memory with the input bytes
+    let mem = GuestMemoryMmap::from_ranges(&[(GuestAddress(0), MEM_SIZE)]).unwrap();
+    if mem.write_slice(mem_bytes, GuestAddress(0 as u64)).is_err() {
+        return;
+    }
+    let guest_memory = GuestMemoryAtomic::new(mem);
+
+    let evt = EventFd::new(0).unwrap();
+    let queue_evt = unsafe { EventFd::from_raw_fd(libc::dup(evt.as_raw_fd())) };
+
+    // Kick the 'queue' event before activate the pmem device
+    queue_evt.write(1).unwrap();
+
+    pmem.activate(
+        guest_memory,
+        Arc::new(NoopVirtioInterrupt {}),
+        vec![(0, q, evt)],
+    )
+    .ok();
+
+    // Wait for the events to finish and pmem device worker thread to return
+    pmem.reset();
+});
+
+fn memfd_create_with_size(name: &ffi::CStr, flags: u32, size: usize) -> Result<RawFd, io::Error> {
+    let fd = unsafe { libc::syscall(libc::SYS_memfd_create, name.as_ptr(), flags) };
+    if fd < 0 {
+        return Err(io::Error::last_os_error());
+    }
+
+    let res = unsafe { libc::syscall(libc::SYS_ftruncate, fd, size) };
+    if res < 0 {
+        return Err(io::Error::last_os_error());
+    }
+
+    Ok(fd as RawFd)
+}
+
+pub struct NoopVirtioInterrupt {}
+
+impl VirtioInterrupt for NoopVirtioInterrupt {
+    fn trigger(&self, _int_type: VirtioInterruptType) -> std::result::Result<(), std::io::Error> {
+        Ok(())
+    }
+}
+
+// Create a dummy virtio-pmem device for fuzzing purpose only
+fn create_dummy_pmem() -> Pmem {
+    let shm =
+        memfd_create_with_size(&ffi::CString::new("fuzz").unwrap(), 0, PMEM_FILE_SIZE).unwrap();
+    let file: File = unsafe { File::from_raw_fd(shm) };
+
+    // The fuzzer is focusing on the virtio-pmem code that processes guest inputs (e.g. virt queues),
+    // so dummy mappings (both mmap and user mapping) does the job and is faster (using smaller amount of memory).
+    let dummy_mapping_size = 1;
+    let cloned_file = file.try_clone().unwrap();
+    let dummy_mmap_region = MmapRegion::build(
+        Some(FileOffset::new(cloned_file, 0)),
+        dummy_mapping_size,
+        PROT_READ | PROT_WRITE,
+        MAP_NORESERVE | MAP_PRIVATE,
+    )
+    .unwrap();
+    let guest_addr = GuestAddress(0);
+    let dummy_user_mapping = UserspaceMapping {
+        host_addr: dummy_mmap_region.as_ptr() as u64,
+        mem_slot: 0,
+        addr: guest_addr,
+        len: dummy_mapping_size as u64,
+        mergeable: false,
+    };
+
+    Pmem::new(
+        "tmp".to_owned(),
+        file,
+        guest_addr,
+        dummy_user_mapping,
+        dummy_mmap_region,
+        false,
+        SeccompAction::Allow,
+        EventFd::new(EFD_NONBLOCK).unwrap(),
+    )
+    .unwrap()
+}
+
+fn setup_virt_queue(bytes: &[u8; QUEUE_DATA_SIZE]) -> Queue {
+    let mut q = Queue::new(QUEUE_SIZE).unwrap();
+    q.set_next_avail(bytes[0] as u16); // 'u8' is enough given the 'QUEUE_SIZE' is small
+    q.set_next_used(bytes[1] as u16);
+    q.set_event_idx(bytes[2] % 2 != 0);
+    q.set_size(bytes[3] as u16 % QUEUE_SIZE);
+
+    q.try_set_desc_table_address(GuestAddress(DESC_TABLE_ADDR))
+        .unwrap();
+    q.try_set_avail_ring_address(GuestAddress(AVAIL_RING_ADDR))
+        .unwrap();
+    q.try_set_used_ring_address(GuestAddress(USED_RING_ADDR))
+        .unwrap();
+    q.set_ready(true);
+
+    q
+}