Abstract: Based on the hourly precipitation observation data of Sichuan Province, the three sets of hourly merged precipitation real-time products with spatial resolutions of 0.01° and 0.05° developed by National Meteorological Information Centre are evaluated and analyzed in the regional rainstorm weather process causing the flood and waterlogging disaster in Baoxing, Ya’an from August 20 to 22, 2019. The results show that the heavy precipitation area and trend of the three sets of merged precipitation real-time products are basically consistent with the actual precipitation, which can reflect the precipitation intensity change and spatiotemporal trends well. The average accuracy rate of the three sets of merged precipitation real-time products is above 85%, and the correlation coefficient is above 0.8, but there are different degrees of underestimaiton compared with the actual precipitation. The applicability of 0.01° product with radar, satellite and site data is the best, 0.05° product merged data of the stations, radars and satellites comes second, 0.05° product only with the stations and satellites is the worst. It can be seen that the fusion site data can improve the quality of the merged precipitation real-time products to a certain extent. In general, the quality of the merged precipitation real-time products is high, which can be used as a supplement to the actual precipitation in the process of disastrous weather.