diff --git a/paddle/fluid/operators/concat_op.cu.cc b/paddle/fluid/operators/concat_op.cu.cc
index 8e38e5231fbf6955ff8a9680a241a4a4ba1b924d..334126c4e0b782c98db2fd3c8278b1daf87da6b6 100644
--- a/paddle/fluid/operators/concat_op.cu.cc
+++ b/paddle/fluid/operators/concat_op.cu.cc
@@ -13,15 +13,20 @@ See the License for the specific language governing permissions and
 limitations under the License. */
 
 #include "paddle/fluid/operators/concat_op.h"
+#include "paddle/fluid/platform/float16.h"
+
 namespace ops = paddle::operators;
+namespace plat = paddle::platform;
 REGISTER_OP_CUDA_KERNEL(
     concat, ops::ConcatKernel<paddle::platform::CUDADeviceContext, double>,
     ops::ConcatKernel<paddle::platform::CUDADeviceContext, float>,
+    ops::ConcatKernel<paddle::platform::CUDADeviceContext, plat::float16>,
     ops::ConcatKernel<paddle::platform::CUDADeviceContext, int64_t>,
     ops::ConcatKernel<paddle::platform::CUDADeviceContext, int>);
 REGISTER_OP_CUDA_KERNEL(
     concat_grad,
     ops::ConcatGradKernel<paddle::platform::CUDADeviceContext, double>,
     ops::ConcatGradKernel<paddle::platform::CUDADeviceContext, float>,
+    ops::ConcatGradKernel<paddle::platform::CUDADeviceContext, plat::float16>,
     ops::ConcatGradKernel<paddle::platform::CUDADeviceContext, int64_t>,
     ops::ConcatGradKernel<paddle::platform::CUDADeviceContext, int>);