To reduce the uncertainty of medium and long-term runoff forecast and increase the power generation efficiency of hydropower reservoirs, a deterministic inflow runoff forecast method based on ensemble Kalman filtering is proposed to address the issue of existing methods focusing on improving the accuracy of deterministic forecast results of a single forecasting model to reduce the uncertainty of runoff forecast.Taking the Jinxi Reservoir as the research object and ten days as the foresight period, conduct a case study.The results show that compared with traditional single forecast models and traditional information fusion forecast models, the medium-and long-term runoff forecast based on ensemble Kalman filtering reduces RMSE by 4.78 m³/s and improves qualification rate by 0.56%. And it effectively reduces the uncertainty of flood season forecasting, obtaining more accurate and reliable deterministic runoff forecasting results, which can provide technical support for the optimization and scheduling of cascade hydropower stations in the basin. Keywords: medium and long-term runoff forecast; data fusion; ensemble Kalman filter; Jinxi Reservoir