diff --git a/torch_npu/csrc/core/npu/NPUCachingAllocator.cpp b/torch_npu/csrc/core/npu/NPUCachingAllocator.cpp
index e3c3a327be49710fd9a9d7ce5540c320a0cc6746..b81750695fe0338b1e55e1c83274aaf676f9804c 100644
--- a/torch_npu/csrc/core/npu/NPUCachingAllocator.cpp
+++ b/torch_npu/csrc/core/npu/NPUCachingAllocator.cpp
@@ -1554,7 +1554,7 @@ public:
             stats.allocated_bytes[static_cast<size_t>(StatType::AGGREGATE)].current,
             stats.reserved_bytes[static_cast<size_t>(StatType::AGGREGATE)].current,
             stats.active_bytes[static_cast<size_t>(StatType::AGGREGATE)].current,
-            reinterpret_cast<int64_t>(block->stream) });
+            block->stream });
 #endif
 
         return block;
@@ -1619,7 +1619,7 @@ public:
             stats.allocated_bytes[static_cast<size_t>(StatType::AGGREGATE)].current,
             stats.reserved_bytes[static_cast<size_t>(StatType::AGGREGATE)].current,
             stats.active_bytes[static_cast<size_t>(StatType::AGGREGATE)].current,
-            reinterpret_cast<int64_t>(block->stream) });
+            block->stream });
 #endif
     }
 
@@ -2434,7 +2434,7 @@ private:
             stats.allocated_bytes[static_cast<size_t>(StatType::AGGREGATE)].current,
             stats.reserved_bytes[static_cast<size_t>(StatType::AGGREGATE)].current,
             stats.active_bytes[static_cast<size_t>(StatType::AGGREGATE)].current,
-            reinterpret_cast<int64_t>(block->stream) });
+            block->stream });
 #endif
     }
 
diff --git a/torch_npu/csrc/core/npu/NPUWorkspaceAllocator.cpp b/torch_npu/csrc/core/npu/NPUWorkspaceAllocator.cpp
index 7d5173dec8d19f2f167609a32c5f02b0c199ae6a..660089b0fb600f6c452f67c89766c98ea9660114 100644
--- a/torch_npu/csrc/core/npu/NPUWorkspaceAllocator.cpp
+++ b/torch_npu/csrc/core/npu/NPUWorkspaceAllocator.cpp
@@ -113,7 +113,7 @@ public:
                     stats.allocated_bytes.current,
                     stats.reserved_bytes.current,
                     stats.allocated_bytes.current,
-                    reinterpret_cast<int64_t>(stream)}
+                    stream }
                 );
 #endif
                 block->data_ptr = nullptr;
@@ -154,7 +154,7 @@ public:
                 stats.allocated_bytes.current,
                 stats.reserved_bytes.current,
                 stats.allocated_bytes.current,
-                reinterpret_cast<int64_t>(stream)}
+                stream }
             );
             this->last_block = block;
             this->last_stream = stream;
@@ -180,7 +180,7 @@ public:
             stats.allocated_bytes.current,
             stats.reserved_bytes.current,
             stats.allocated_bytes.current,
-            reinterpret_cast<int64_t>(stream)}
+            stream }
         );
         this->last_block = block;
         this->last_stream = stream;
@@ -204,7 +204,7 @@ public:
                 stats.allocated_bytes.current,
                 stats.reserved_bytes.current,
                 stats.allocated_bytes.current,
-                reinterpret_cast<int64_t>(this->last_stream)}
+                this->last_stream }
             );
         }
 #endif
@@ -254,7 +254,7 @@ public:
                     stats.allocated_bytes.current,
                     stats.reserved_bytes.current,
                     stats.allocated_bytes.current,
-                    reinterpret_cast<int64_t>(block_pair.first)}
+                    block_pair.first }
                 );
 #endif
             }
diff --git a/torch_npu/csrc/core/npu/interface/AclInterface.cpp b/torch_npu/csrc/core/npu/interface/AclInterface.cpp
index 30a4280edc8ccb71e202f57ba1901010fd241c49..ef5e7eeb13b7754245ec24318389c66730324991 100644
--- a/torch_npu/csrc/core/npu/interface/AclInterface.cpp
+++ b/torch_npu/csrc/core/npu/interface/AclInterface.cpp
@@ -89,6 +89,7 @@ LOAD_FUNCTION(aclrtIpcMemClose)
 LOAD_FUNCTION(aclrtMemExportToShareableHandle)
 LOAD_FUNCTION(aclrtMemSetPidToShareableHandle)
 LOAD_FUNCTION(aclrtMemImportFromShareableHandle)
+LOAD_FUNCTION(aclrtStreamGetId)
 LOAD_FUNCTION(aclrtDeviceGetBareTgid)
 
 aclprofStepInfoPtr init_stepinfo() {
@@ -1033,5 +1034,16 @@ aclError AclrtDeviceGetBareTgid(int32_t *pid)
     return func(pid);
 }
 
+aclError AclrtStreamGetId(aclrtStream stream, int32_t* stream_id)
+{
+    typedef aclError(*AclrtStreamGetIdFunc)(aclrtStream, int32_t*);
+    static AclrtStreamGetIdFunc func = nullptr;
+    if (func == nullptr) {
+        func = (AclrtStreamGetIdFunc)GET_FUNC(aclrtStreamGetId);
+    }
+    TORCH_CHECK(func, "Failed to find function ", "AclrtStreamGetId", PROF_ERROR(ErrCode::NOT_FOUND));
+    return func(stream, stream_id);
+}
+
 } // namespace acl
 } // namespace c10
diff --git a/torch_npu/csrc/core/npu/interface/AclInterface.h b/torch_npu/csrc/core/npu/interface/AclInterface.h
index 373aca671f5e5e0a73de73c72826c5c17c229c4f..46dd1180200b0ab598efad49c84251abf2cd0843 100644
--- a/torch_npu/csrc/core/npu/interface/AclInterface.h
+++ b/torch_npu/csrc/core/npu/interface/AclInterface.h
@@ -243,6 +243,8 @@ aclError AclrtMemSetPidToShareableHandle(uint64_t shareableHandle, int32_t *pid,
 
 aclError AclrtMemImportFromShareableHandle(uint64_t shareableHandle, int32_t deviceId, aclrtDrvMemHandle *handle);
 
+aclError AclrtStreamGetId(aclrtStream stream, int32_t* stream_id);
+
 aclError AclrtDeviceGetBareTgid(int32_t *pid);
 
 } // namespace acl
diff --git a/torch_npu/csrc/profiler/npu_profiler.cpp b/torch_npu/csrc/profiler/npu_profiler.cpp
index 295eda9aea1f2a425c21caaf037e4f60713a463e..3678da07550d017db863d53ff2d7d05cf12f4edd 100644
--- a/torch_npu/csrc/profiler/npu_profiler.cpp
+++ b/torch_npu/csrc/profiler/npu_profiler.cpp
@@ -6,6 +6,7 @@
 
 #include "torch_npu/csrc/core/npu/npu_log.h"
 #include "torch_npu/csrc/core/npu/NPUException.h"
+#include "torch_npu/csrc/core/npu/interface/AclInterface.h"
 #include "torch_npu/csrc/profiler/npu_profiler.h"
 
 #include "torch_npu/csrc/toolkit/profiler/common/utils.h"
@@ -380,6 +381,8 @@ void reportMemoryDataToNpuProfiler(const MemoryUsage& data)
     if (!ProfilerMgr::GetInstance()->ReportMemEnable().load()) {
         return;
     }
+    int32_t stream_id;
+    c10_npu::acl::AclrtStreamGetId(data.stream, &stream_id);
     ProfilerMgr::GetInstance()->UploadWithLock(std::make_unique<torch_npu::toolkit::profiler::MemoryData>(
         data.ptr,
         static_cast<int64_t>(Utils::GetClockTime()),
@@ -387,7 +390,7 @@ void reportMemoryDataToNpuProfiler(const MemoryUsage& data)
         data.total_allocated,
         data.total_reserved,
         data.total_active,
-        data.stream_ptr,
+        stream_id,
         data.device_type,
         data.device_index,
         data.component_type,
diff --git a/torch_npu/csrc/profiler/npu_profiler.h b/torch_npu/csrc/profiler/npu_profiler.h
index 2127825bc134e0b49178fe00890cdff58011e62c..715660f2e9fd90191e1a0061ab090f7599f730a0 100644
--- a/torch_npu/csrc/profiler/npu_profiler.h
+++ b/torch_npu/csrc/profiler/npu_profiler.h
@@ -7,6 +7,8 @@
 
 #include <ATen/record_function.h>
 
+#include "third_party/acl/inc/acl/acl_base.h"
+#include "third_party/acl/inc/acl/acl_rt.h"
 #include "torch_npu/csrc/toolkit/profiler/inc/data_reporter.h"
 #include "torch_npu/csrc/profiler/profiler_mgr.h"
 #include "torch_npu/csrc/profiler/mstx_mgr.h"
@@ -55,7 +57,7 @@ struct MemoryUsage {
     int64_t total_allocated{0};
     int64_t total_reserved{0};
     int64_t total_active{0};
-    int64_t stream_ptr{0};
+    aclrtStream stream{nullptr};
 };
 
 struct ExperimentalConfig {