Water resources are the source of life and the basis of ecology, and medium-and long-term runoff forecasting plays an important role in the overall planning of water resources. However, complex factors such as climate change affect the formation of runoff, making the runoff process more complex and posing a great challenge to hydrological forecasting. Therefore, this paper investigates the medium-and long-term runoff forecasting in the Huaihe River basin based on machine learning. The paper takes the improvement of medium-and long-term runoff prediction accuracy as the background, and combines the specific requirements of practical projects and research ideas to characterise the runoff series in the Huaihe River basin. The results of runoff simulation and forecasting under future climate change are also analyzed using the BOA-EEMD-LSTM model.