Abstract: The black soil region in the cold temperate zone of Northeast China is a vital grain production area in the country, and water resources are crucial for ensuring food security.Evaporation is a key process in understanding the water cycle, so enhancing evaporation prediction capabilities is essential for understanding water resource changes in Northeast China and ensuring agricultural development and food security.This study focuses on the western region of Heilongjiang Province in Northeast China as a representative area and proposes a prediction model combining Empirical Mode Decomposition (EMD) and Support Vector Machine (SVM), referred to as the EMD-SVM model, to address the evaporation prediction challenges in the cold temperate black soil region of Northeast China.The performance of the proposed model was compared with traditional SVM and EMD-ARIMA models.The results show that the EMD-SVM model outperforms the traditional SVM and EMD-ARIMA models in both the training and validation periods, particularly in fitting low evaporation sequences.During the validation period, the EMD-SVM model achieved a Nash-Sutcliffe Efficiency (NSE) of 0.88 and a correlation coefficient (R) of 0.95, while the Mean Absolute Error (MAE) was reduced by 32.7% compared to the SVM model.This study provides a novel methodological reference for improving evaporation prediction capabilities in the cold temperate black soil region of Northeast China.