Based on the raindrop spectrum observation data of Urumqi, Xinjiang from July 3 to October 3, 2018, this paper improved the WRF Single-Moment 6-class (WSM6) scheme in Urumqi regional high-resolution numerical prediction system, and evaluated the improvement effect of a heavy precipitation process prediction in Xinjiang from 1200 BJT on June 15 to 0000 BJT on June 17, 2021. The results show that the average diameter (D0), maximum diameter (Dmax), and mass weighted average diameter (Dm) of raindrops in Urumqi were 0.65 mm, 1.60 mm, and 0.93 mm, respectively based on the raindrop spectrum observation. The WSM6-new scheme considering the fitting relationship between parameters lgNw and Dm in Xinjiang improved the prediction ability of the intensity and strong center range of precipitation to a certain extent. The scoring results of TS, BR, ETS, and TSS show that with the increase of precipitation grade, the prediction ability of WSM6-new scheme was significantly improved, and it showed obvious advantages for the prediction of heavy rain and torrential rain. The parameter scheme with different raindrop size distribution had a certain impact on the structural characteristics of precipitation cloud system, vertical velocity, atmospheric stratification, and divergence field. The impact on cloud microphysical processes was mainly reflected in the content and distribution of rainwater. WSM6-new scheme adopted the statistical characteristics of raindrop spectrum in Xinjiang, which made the description of raindrop size distribution in the model closer to reality. The number concentration of large raindrops in raindrop spectrum increased obviously, the terminal velocity of raindrops increased, and the drag effect enhanced, which was conducive to the enhancement and maintenance of downdraft below frozen level. The strong downdraft formed a strong divergent outflow in the near ground layer, which strengthened the air convergence in the convection area near the ground and was of benefit to the development and strengthening of the updraft. Therefore, a stronger precipitation process appeared on the ground, and the prediction ability for heavy rain and torrential rain was obviously enhanced.