Frequent heavy snowfall in Xinjiang plays an important role in the land water cycle. In this study, 18 groups of simulation experiments are conducted on the heavy snowfall event in Xinjiang during 9–13 December of 2015 using the Weather Research and Forecasting (WRF) model. In these experiments, the combination of six land surface parameterization schemes (the Noah scheme, Noah-MP scheme, RUC scheme, CLM4 scheme, PX scheme, and TD scheme) with three microphysical parameterization schemes (the WSM6 scheme, Thompson scheme, and Lin scheme) are adopted, where the observed snowfall data are used for performance evaluation. Results show that the simulated snowfall intensity and snowfall range in different areas are very sensitive to the selection of the land surface scheme. The snowfall in southern Xinjiang is overestimated by almost all six schemes, where the Noah-MP scheme performs more reasonably than the others. The Noah scheme shows its advantage in northwestern Xinjiang. The three different microphysical schemes vary significantly in producing snowfall amount. The WSM6 scheme produced the largest snowfall amount, and the Lin scheme resulted in the smallest snowfall amount. In addition, the accumulated snowfall amounts above 10 mm are generally underestimated by all six land surface schemes, while the accumulated snowfall amounts below 10 mm are overestimated by most of the schemes. The Noah-MP scheme performs the best in the simulation of the snowfall amount in the whole region. However, the Noah scheme shows an advantage in areas with a large snowfall amount.