diff --git a/paddle/fluid/operators/cast_op.cc b/paddle/fluid/operators/cast_op.cc
index 952e9ca329f102566d14cbf9180001e4ae5aef35..5fc97924ef27fedd7503207ea6e8cd7d63ecb1a2 100644
--- a/paddle/fluid/operators/cast_op.cc
+++ b/paddle/fluid/operators/cast_op.cc
@@ -119,6 +119,7 @@ REGISTER_OPERATOR(cast, ops::CastOp,
 REGISTER_OP_CPU_KERNEL(
     cast, ops::CastOpKernel<CPU, float>, ops::CastOpKernel<CPU, double>,
     ops::CastOpKernel<CPU, int>, ops::CastOpKernel<CPU, int64_t>,
+    ops::CastOpKernel<CPU, int>, ops::CastOpKernel<CPU, int16_t>,
     ops::CastOpKernel<CPU, bool>, ops::CastOpKernel<CPU, uint8_t>,
     ops::CastOpKernel<CPU, paddle::platform::float16>,
     ops::CastOpKernel<CPU, paddle::platform::bfloat16>,
diff --git a/paddle/fluid/operators/cast_op.cu b/paddle/fluid/operators/cast_op.cu
index 0beb2291060169628a7d736a21d728993a2bc0e5..8b2998e52a172e738c9dd59983729be32357b326 100644
--- a/paddle/fluid/operators/cast_op.cu
+++ b/paddle/fluid/operators/cast_op.cu
@@ -102,6 +102,7 @@ REGISTER_OP_CUDA_KERNEL(
     ops::CastOpKernel<paddle::platform::CUDADeviceContext, double>,
     ops::CastOpKernel<paddle::platform::CUDADeviceContext, int>,
     ops::CastOpKernel<paddle::platform::CUDADeviceContext, int64_t>,
+    ops::CastOpKernel<paddle::platform::CUDADeviceContext, int16_t>,
     ops::CastOpKernel<paddle::platform::CUDADeviceContext, bool>,
     ops::CastOpKernel<paddle::platform::CUDADeviceContext, uint8_t>,
     ops::CastOpKernel<paddle::platform::CUDADeviceContext,
@@ -116,6 +117,7 @@ REGISTER_OP_CUDA_KERNEL(
     ops::CastOpKernel<paddle::platform::CUDADeviceContext, double>,
     ops::CastOpKernel<paddle::platform::CUDADeviceContext, int>,
     ops::CastOpKernel<paddle::platform::CUDADeviceContext, int64_t>,
+    ops::CastOpKernel<paddle::platform::CUDADeviceContext, int16_t>,
     ops::CastOpKernel<paddle::platform::CUDADeviceContext, bool>,
     ops::CastOpKernel<paddle::platform::CUDADeviceContext, uint8_t>,
     ops::CastOpKernel<paddle::platform::CUDADeviceContext,