Abstract: Selecting the precipitation process in Yunnan on July 23, 2019, introducing the C-band Doppler radar reflectivity factor data into the CMA-GD numerical model through a cloud analysis system, the impact of radar reflectivity factor assimilation on the model prediciton effect was analyzed. The results show that: (1) Introducing the reflectivity factor could significantly improve the prediction of precipitation intensity within 3 hours, but the improvement on precipitation areas was not significant. (2) Cloud water, graupel, raindrops, and cloud snow increased significantly in the first 6 hours, while cloud ice and water vapor continued to increase until the 16th hour. The most significant increases were observed: cloud ice at 300 hPa, cloud snow at 400 hPa, graupel at 500 hPa, water vapor at 600 hPa, and cloud water and raindrops at 700 hPa. (3) The relative humidity increased by 2% to 3% in the first 12 hours at 600~500 hPa, and from the 12th to 21st hour it creased by 2% to 3% at 300~100 hPa. The relative humidity decreased in the first 9 hours below 850 hPa, and decreased by 5% in the first 3 hours. Similar to water condensate, the spatial correspondence between the high value area of precipitation in the first 3 hours and the high value area of water condensate and relative humidity was not very consistent, which might be due to the fact that the water vapor condition was only caused by the background effect during the precipitation process. (4) In the southwest and east-southeast of Yunnan with strong precipitation, the wind field showed high-level divergence and low-level convergence after introducing the reflectivity factor, which was conducive to the occurrence of precipitation. The analysis of water condensate, relative humidity, and wind field revealed that the precipitation area in this event was mainly determined by local lupifting effects.