diff --git a/mindspore/ccsrc/kernel/tbe/tbe_kernel_build.cc b/mindspore/ccsrc/kernel/tbe/tbe_kernel_build.cc
index 7d706eec66..585747d303 100644
--- a/mindspore/ccsrc/kernel/tbe/tbe_kernel_build.cc
+++ b/mindspore/ccsrc/kernel/tbe/tbe_kernel_build.cc
@@ -721,7 +721,7 @@ size_t TbeKernelBuild::GetOptionalInput(const mindspore::CNodePtr &cnode, bool i
   return (op_info->inputs_ptr().size() + 1 - cnode->inputs().size());
 }
 
-std::string TbeKernelBuild::GetRealOpType(std::string &origin_type) {
+std::string TbeKernelBuild::GetRealOpType(const std::string &origin_type) {
   static std::map<std::string, std::string> buffer_fussion_op_map = {{"DepthwiseConv2dNative", "DepthwiseConv2D"},
                                                                      {"TensorAdd", "Add"}};
   string result = origin_type;
@@ -834,9 +834,9 @@ bool TbeKernelBuild::GenFusionComputeJson(const mindspore::AnfNodePtr &compute_n
   }
   (*compute_op_str)["output_desc"] = output_desc_list;
   // gen others
-  auto type = AnfAlgo::GetCNodeName(cnode);
+  auto origin_type = AnfAlgo::GetCNodeName(cnode);
   // replace special op type for buffer fusion op
-  type = GetRealOpType(type);
+  auto type = GetRealOpType(origin_type);
   (*compute_op_str)["type"] = type;
   tbe::TbeAdapter::NormalizeFuncName(&type);
   (*compute_op_str)["func_name"] = type;
diff --git a/mindspore/ccsrc/kernel/tbe/tbe_kernel_build.h b/mindspore/ccsrc/kernel/tbe/tbe_kernel_build.h
index c08d88fb35..2ddab34d49 100644
--- a/mindspore/ccsrc/kernel/tbe/tbe_kernel_build.h
+++ b/mindspore/ccsrc/kernel/tbe/tbe_kernel_build.h
@@ -76,7 +76,7 @@ class TbeKernelBuild {
                              std::map<const AnfNodePtr, FusionDataType> *spec_data_input);
   static bool IsDynamicInput(const CNodePtr &cnode);
   static size_t GetOptionalInput(const CNodePtr &cnode, bool is_dynamic_input);
-  std::string GetRealOpType(std::string &origin_type);
+  static std::string GetRealOpType(const std::string &origin_type);
 };
 
 class TbeKernelJsonCreator {