Abstract: By using SRTM digital elevation DEM model and land use data, we took ArcGIS as the platform and Qijiang River Basin as the research object, and adopted five different types of precipitation data such as automatic weather station precipitation data and multi-source precipitation fusion products to calculate the hourly basin area rainfall with Tyson polygon method.The FloodArea hydrodynamic model was driven to simulate the water depth of a mountain flood process in Qijiang River Basin from 16:00 8 August to 10:00 9 August 2021.And then the simulated water depth and the measured water level were compared and analyzed.The study shows that 1) the correlation coefficients between the simulated depth and the measured depth from five types of precipitation data all reach a relatively high level, with 96% as high as 0.93.The highest correlation coefficient of 0.9857 can be found in AWS, RT and NRT data; 2) On the whole, the simulated water level is basically consistent with the measured water level, but on the peak, the simulated water level has a lag of 1-2 hours, with slightly lower peak level from the simulated data; 3) The mountain torrent processes simulated by five types of precipitation data exhibit high fitting degree, and all of the certainty coefficients are equal to or higher than 0.8715, with 96% as high as 0.9215 and 0.9831 as the highest; 4) The mountain torrents simulated by five types of precipitation data have high reliability, and the Nash efficiency coefficients exceed 0.8687, with 96%as high as 0.9202 and 0.9817 as the highest.