diff --git a/hydromodel/trainers/evaluate.py b/hydromodel/trainers/evaluate.py
index 70421f0..49e5cb7 100644
--- a/hydromodel/trainers/evaluate.py
+++ b/hydromodel/trainers/evaluate.py
@@ -1,7 +1,7 @@
 """
 Author: Wenyu Ouyang
 Date: 2022-10-25 21:16:22
-LastEditTime: 2024-09-17 14:23:29
+LastEditTime: 2024-09-17 15:13:57
 LastEditors: Wenyu Ouyang
 Description: Plots for calibration and testing results
 FilePath: \hydromodel\hydromodel\trainers\evaluate.py
@@ -325,22 +325,25 @@ def _read_save_sceua_calibrated_params(basin_id, save_dir, sceua_calibrated_file
     # the following code is from spotpy but its performance is not good so we use pandas to replace it
     # results = spotpy.analyser.load_csv_results(sceua_calibrated_file_name)
     # bestindex, bestobjf = spotpy.analyser.get_minlikeindex(results)
-    best_model_run = results[bestindex]
-    fields = [word for word in best_model_run.dtype.names if word.startswith("par")]
-    best_calibrate_params = pd.DataFrame(list(best_model_run[fields]))
+    best_model_run = results.iloc[bestindex]
+    fields = [word for word in best_model_run.index if word.startswith("par")]
+    best_calibrate_params = pd.DataFrame(
+        [best_model_run[fields].values], columns=fields
+    )
     save_file = os.path.join(save_dir, basin_id + "_calibrate_params.txt")
-    best_calibrate_params.to_csv(save_file, sep=",", index=False, header=True)
+    # to keep consistent with the original code, we save the best parameters to a txt file
+    best_calibrate_params.T.to_csv(save_file, index=False, columns=None)
     # Return the best result as a single row
-    return np.array(best_calibrate_params).reshape(1, -1)
+    return best_calibrate_params.to_numpy().reshape(1, -1)
 
 
 def _read_all_basin_params(basins, param_dir):
     params_list = []
     for basin_id in basins:
         db_name = os.path.join(param_dir, basin_id)
-        # 读取每个流域的参数
+        # Read parameters for each basin
         basin_params = _read_save_sceua_calibrated_params(basin_id, param_dir, db_name)
-        # 确保basin_params是一维的
+        # Ensure basin_params is one-dimensional
         basin_params = basin_params.flatten()
         params_list.append(basin_params)
     return np.vstack(params_list)