Taking the Yellow River source area as the study area, a comprehensive evaluation was conducted on the multi-source fusion precipitation product(DMLTPP), and it was compared with the pure infrared rainfall product PERSIANNCCS, satellite precipitation product GSMaPG, and reanalysis dataset ERA5L. The results show that ERA5L overestimates precipitation to a certain extent, but reflects the spatiotemporal pattern of precipitation in the entire Yellow River source area well. GSMaP_G performs better in precipitation estimation compared to other products. In the scheme of calibrating the parameters of the grid Xin’anjiang model with interpolated data from meteorological stations, GSMaP_G has a better evaluation of hydrological effects in the Menma section of the Yellow River source area. In the scheme of using precipitation products to calibrate the parameters of the grid Xin’anjiang model, the runoff simulation based on DMLTPP and ERA5L performed the best in normal years, while the simulation results based on GSMaP_G better reproduced the flood peak in wet years.