diff --git a/src/plugins/intel_cpu/src/nodes/executors/acl/acl_deconv.cpp b/src/plugins/intel_cpu/src/nodes/executors/acl/acl_deconv.cpp
index fa40b3a27322c2..ada33e055777c0 100644
--- a/src/plugins/intel_cpu/src/nodes/executors/acl/acl_deconv.cpp
+++ b/src/plugins/intel_cpu/src/nodes/executors/acl/acl_deconv.cpp
@@ -91,13 +91,23 @@ bool AclDeconvExecutor::init(const DeconvAttrs& deconvAttrs,
         return false;
     }
 
+    lifetime_mgr = std::make_shared<arm_compute::BlobLifetimeManager>();
+    pool_mgr = std::make_shared<arm_compute::PoolManager>();
+    aclMemoryManager = std::make_shared<arm_compute::MemoryManagerOnDemand>(lifetime_mgr, pool_mgr);
+    aclMemoryGroup = std::make_shared<arm_compute::MemoryGroup>(aclMemoryManager);
+
+    aclMemoryGroup->manage(&srcTensor);
+    aclMemoryGroup->manage(&weiTensor);
+    aclMemoryGroup->manage(&dstTensor);
     srcTensor.allocator()->init(srcTensorInfo);
     weiTensor.allocator()->init(weiTensorInfo);
     dstTensor.allocator()->init(dstTensorInfo);
-    if (deconvAttrs.withBiasesParam)
+    if (deconvAttrs.withBiasesParam) {
+        aclMemoryGroup->manage(&biasTensor);
         biasTensor.allocator()->init(biasTensorInfo);
+    }
 
-    deconv = std::make_unique<arm_compute::NEDeconvolutionLayer>();
+    deconv = std::make_unique<arm_compute::NEDeconvolutionLayer>(aclMemoryManager);
     configureThreadSafe([&] {
         deconv->configure(&srcTensor, &weiTensor, deconvAttrs.withBiasesParam ? &biasTensor : nullptr, &dstTensor, deconv_info, deconvAttrs.aclFastMath);
     });
@@ -185,6 +195,11 @@ void AclDeconvExecutor::exec(const std::vector<MemoryCPtr>& src, const std::vect
     dstTensor.allocator()->import_memory(dst[0]->getData());
     if (deconvAttrs.withBiasesParam)
         biasTensor.allocator()->import_memory(src[2]->getData());
+
+    arm_compute::Allocator allocator{};
+    aclMemoryManager->populate(allocator, 1);
+    aclMemoryGroup->acquire();
+
     deconv->run();
 
     srcTensor.allocator()->free();
@@ -192,6 +207,9 @@ void AclDeconvExecutor::exec(const std::vector<MemoryCPtr>& src, const std::vect
     weiTensor.allocator()->free();
     if (deconvAttrs.withBiasesParam)
         biasTensor.allocator()->free();
+
+    aclMemoryGroup->release();
+    aclMemoryManager->clear();
 }
 
 bool AclDeconvExecutorBuilder::customIsSupported(const DeconvAttrs &deconvAttrs,
diff --git a/src/plugins/intel_cpu/src/nodes/executors/acl/acl_deconv.hpp b/src/plugins/intel_cpu/src/nodes/executors/acl/acl_deconv.hpp
index ad743690f45c52..6d4750319c0559 100644
--- a/src/plugins/intel_cpu/src/nodes/executors/acl/acl_deconv.hpp
+++ b/src/plugins/intel_cpu/src/nodes/executors/acl/acl_deconv.hpp
@@ -6,6 +6,10 @@
 
 #include "nodes/executors/deconv.hpp"
 #include "arm_compute/runtime/NEON/NEFunctions.h"
+#include "arm_compute/runtime/PoolManager.h"
+#include "arm_compute/runtime/MemoryManagerOnDemand.h"
+#include "arm_compute/runtime/BlobLifetimeManager.h"
+#include "arm_compute/runtime/Allocator.h"
 #include "utils/debug_capabilities.h"
 #include "acl_utils.hpp"
 #include "src/cpu/CpuTypes.h"
@@ -48,6 +52,10 @@ class AclDeconvExecutor : public DeconvExecutor {
     arm_compute::Tensor weiTensor;
     arm_compute::Tensor biasTensor;
     arm_compute::Tensor dstTensor;
+    std::shared_ptr<arm_compute::BlobLifetimeManager> lifetime_mgr = nullptr;
+    std::shared_ptr<arm_compute::PoolManager> pool_mgr = nullptr;
+    std::shared_ptr<arm_compute::MemoryManagerOnDemand> aclMemoryManager = nullptr;
+    std::shared_ptr<arm_compute::MemoryGroup> aclMemoryGroup = nullptr;
     std::unique_ptr<arm_compute::NEDeconvolutionLayer> deconv = nullptr;
 };
 
diff --git a/src/plugins/intel_cpu/thirdparty/ComputeLibrary b/src/plugins/intel_cpu/thirdparty/ComputeLibrary
index c61bd3387403b7..f44f09dfd263c2 160000
--- a/src/plugins/intel_cpu/thirdparty/ComputeLibrary
+++ b/src/plugins/intel_cpu/thirdparty/ComputeLibrary
@@ -1 +1 @@
-Subproject commit c61bd3387403b76d618915ccebf5e9585f52a071
+Subproject commit f44f09dfd263c2bb0238f4323e366a210e9d4aae