Abstract: Based on the precipitation data from the meteorological and hydrological stations and the Quantitative Precipitation Estimation (QPE) from radars over Jiangxi between May and June in 2019, we have made a fusion test of hourly precipitation from the meteorological and hydrological stations and by radars with 1 km×1 km resolution by using the space-time multiscale analysis system (STMAS), and made an analysis and assessment of fusion test results. The results show that the precipitation detected by meteorological stations, hydrological stations and radars has a similar spatial distribution. It is found that the severe precipitation zones and the patterns of precipitation distribution are very consistent. Comparing with the precipitation observed, the different fusion test results have a consistent variation tendency, but all are the underestimating to a certain extent. Those fused precipitation from the hydrological stations have a significant improvement in underestimation. The independent verification results indicate that the test results fused precipitation from hydrological stations have an obvious improvement in the accuracy of different level of precipitation, whose correlation coefficient increases by 9.2%, and root mean square error and mean absolute error decreases by 14.3% and 12.1%, respectively. After further fusing precipitation from hydrological stations, precipitation fusion results of different height areas are significantly improved, especially in the areas of elevation greater than 800 m whose correlation coefficient increases by 8.8% and root mean square error decreases by 14.1%.