Based on the surface precipitation observation data and the four merged precipitation products (FAST_5 km , FRT_5 km, RT_1 km, NRT_1 km) developed by National Meteorological Information Center, analysis was performed for the applicability of the merged precipitation products in western Sichuan Basin by using error analysis method, variance analysis method, accuracy method and TS score method. Detailed evaluation was performed for an extreme rainstorm process occurred during 10th Aug to 11th Aug in 2020. The results show merged precipitation products was nearly consistent with the observation data, and the 1 km product was closer to the observation data. For the accumulative precipitation of the extreme rainstorm process, the four merged products could reflect the precipitation process well, and the precipitation area, trend and rain band pattern were consistent with the observed data. The precipitation intensity and area from 1 km product was closer to the observed data, and the NRT_1 km had the highest matching degree and smallest deviation compared to the observation data. The 24 h rainfall maximum value of merged precipitation products was smaller than observation data, and the maximum value from 1 km product had a big improvement compared to 5 km product. The hourly maximum precipitation of merged precipitation products was smaller than observation data, and there was a certain amount of deviation. During the strong precipitation period, merged precipitation products were nearly consistent with the observation data, and could reflect the greater value area of the precipitation process. On the whole, the effect of three-source merged grid products were better than that of two-source merged grid products, and the near real-time three-source merged grid product integrating CMORPH and FY satellite data was the best.