Abstract: To improve the prediction period and forecasting accuracy of rainfall in the basin, and enhance the level of flood control, disaster reduction and emergency response in the basin, taking Nenjiang Basin as an example, multi-source meteorological data were adopted to conduct rainfall forecasting for the target basin. The fusion method of multi-source meteorological data was constructed. The WRF model of the basin was built and the rainfall forecast correction based on global meteorological forcing was carried out. The differences between WRF model and the GFS forecast were compared in terms of rainfall forecast accuracy and forecast period. The research results showed that by optimizing the physical parameterization scheme of the rainfall model, the accuracy and reliability of rainfall forecasting in the studied basin can be significantly improved. The method is applicable to the basin of this study, which shows higher precision in capturing the physical process of rainfall, and can better simulate the temporal and spatial distribution characteristics of rainfall. The research results have certain practical significance for rainfall forecasting and flood disaster prevention in the river basin.