diff --git a/src/plugins/intel_gpu/include/intel_gpu/runtime/debug_configuration.hpp b/src/plugins/intel_gpu/include/intel_gpu/runtime/debug_configuration.hpp
index ea632ee4ac5..febf41a557f 100644
--- a/src/plugins/intel_gpu/include/intel_gpu/runtime/debug_configuration.hpp
+++ b/src/plugins/intel_gpu/include/intel_gpu/runtime/debug_configuration.hpp
@@ -121,6 +121,7 @@ public:
     std::vector<std::string> forced_impl_types;                 // Force implementation type either ocl or onednn
     int max_kernels_per_batch;                                  // Maximum number of kernels in a batch during compiling kernels
     int disable_async_compilation;                              // Disable async compilation
+    int disable_winograd_conv;                                  // Disable Winograd conv
     int disable_dynamic_impl;                                   // Disable dynamic implementation
     int disable_runtime_buffer_fusing;                          // Disable runtime buffer fusing
     int disable_memory_reuse;                                   // Disable memmory reuse among layers
diff --git a/src/plugins/intel_gpu/src/graph/layout_optimizer.cpp b/src/plugins/intel_gpu/src/graph/layout_optimizer.cpp
index 00f75f37fb7..005b9a21823 100644
--- a/src/plugins/intel_gpu/src/graph/layout_optimizer.cpp
+++ b/src/plugins/intel_gpu/src/graph/layout_optimizer.cpp
@@ -505,6 +505,10 @@ bool should_use_winograd_2x3_s1(const convolution_node& node,
         return false;
 
     // cases when NOT to use winograd
+    GPU_DEBUG_GET_INSTANCE(debug_config);
+    GPU_DEBUG_IF(debug_config->disable_winograd_conv == 1)
+        return false;
+
     auto prim = node.get_primitive();
     if (input_layout.data_type != data_types::f16
         || input_layout.feature() % 64 != 0  // current algorithm is effective for ifm to be multiply of 64
diff --git a/src/plugins/intel_gpu/src/runtime/debug_configuration.cpp b/src/plugins/intel_gpu/src/runtime/debug_configuration.cpp
index 4e83f109241..084d0095dd2 100644
--- a/src/plugins/intel_gpu/src/runtime/debug_configuration.cpp
+++ b/src/plugins/intel_gpu/src/runtime/debug_configuration.cpp
@@ -134,9 +134,12 @@ static void print_help_messages() {
                               " For primitives fc, gemm, do, reduce, concat are supported. Separated by space.");
     message_list.emplace_back("OV_GPU_MaxKernelsPerBatch", "Maximum number of kernels in a batch during compiling kernels");
     message_list.emplace_back("OV_GPU_DisableAsyncCompilation", "Disable async compilation");
+    message_list.emplace_back("OV_GPU_DisableWinogradConv", "Disable Winograd convolution");
     message_list.emplace_back("OV_GPU_DisableDynamicImpl", "Disable dynamic implementation");
     message_list.emplace_back("OV_GPU_DisableRuntimeBufferFusing", "Disable runtime buffer fusing");
     message_list.emplace_back("OV_GPU_DisableMemoryReuse", "Disable memory reuse");
+    message_list.emplace_back("OV_GPU_DisableBuildTimeWeightReorderForDynamicNodes", "Disable build time weight reorder for dynmaic nodes.");
+    message_list.emplace_back("OV_GPU_DisableRuntimeSkipReorder", "Disable runtime skip reorder.");
     message_list.emplace_back("OV_GPU_DisablePrimitiveFusing", "Disable primitive fusing");
     message_list.emplace_back("OV_GPU_DumpIteration", "Dump n-th execution of network, separated by space.");
     message_list.emplace_back("OV_GPU_MemPreallocationOptions", "Controls buffer pre-allocation feature. Expects 4 values separated by space in"
@@ -188,6 +191,7 @@ debug_configuration::debug_configuration()
         , serialize_compile(0)
         , max_kernels_per_batch(0)
         , disable_async_compilation(0)
+        , disable_winograd_conv(0)
         , disable_dynamic_impl(0)
         , disable_runtime_buffer_fusing(0)
         , disable_memory_reuse(0)
@@ -225,6 +229,7 @@ debug_configuration::debug_configuration()
     get_gpu_debug_env_var("ForceImplTypes", forced_impl_types_str);
     get_gpu_debug_env_var("MaxKernelsPerBatch", max_kernels_per_batch);
     get_gpu_debug_env_var("DisableAsyncCompilation", disable_async_compilation);
+    get_gpu_debug_env_var("DisableWinogradConv", disable_winograd_conv);
     get_gpu_debug_env_var("DisableDynamicImpl", disable_dynamic_impl);
     get_gpu_debug_env_var("DisableRuntimeBufferFusing", disable_runtime_buffer_fusing);
     get_gpu_debug_env_var("DisableMemoryReuse", disable_memory_reuse);