diff --git a/src/transformers/trainer_utils.py b/src/transformers/trainer_utils.py
index d17113091777be..a52560a77cf339 100644
--- a/src/transformers/trainer_utils.py
+++ b/src/transformers/trainer_utils.py
@@ -71,6 +71,9 @@ def enable_full_determinism(seed: int, warn_only: bool = False):
         # depending on the CUDA version, so we set them both here
         os.environ["CUDA_LAUNCH_BLOCKING"] = "1"
         os.environ["CUBLAS_WORKSPACE_CONFIG"] = ":16:8"
+        # The environment variable required to enable deterministic mode on Ascend NPUs.
+        os.environ["ASCEND_LAUNCH_BLOCKING"] = 1
+        os.environ["HCCL_DETERMINISTIC"] = 1
         torch.use_deterministic_algorithms(True, warn_only=warn_only)
 
         # Enable CUDNN deterministic mode