Eight different cloud microphysical schemes in the WRF model are used to simulate the precipitation process caused by a squall line in Hainan Island on 22 April 2020, and the effects of different cloud microphysical schemes on the simulation of the Hainan Island squall line are comparatively analyzed. The results show that different cloud microphysical schemes have significant differences in simulating the surface precipitation, radar composite reflectivity, thermodynamic and dynamic fields. Among them, the precipitation area and center intensity simulated by Thompson scheme are most close to the actual observations, and the radar composite reflectivity in intensity, range and pattern simulated by WSM6 scheme at the time of heaviest precipitation is similar to the actual observations. In the thermodynamic and dynamic fields, the characteristics of squall line such as surface cold pool, low-level vertical wind shear and cold pool outflow can be simulated by all schemes, and the precipitation center corresponds to the strong updraft zone. The divergence structure of low-level convergence and high-level divergence is conducive to the occurrence of severe convection and the formation of precipitation, but there exist differences in the intensity and distribution of precipitation. According to the cloud microphysical characteristics, the liquid-phase particles are mainly distributed below 5 km, and the ice-phase particales are above 6 km. The simulation results of cloud water show the weakest response to the selection of cloud microphysical schemes, while the distributions of snow and graupel show the high sensitivity. This is because different cloud microphysical schemes have different processing ways for the generation, transformation and consumption of snow and graupel, and the conversion rate of the same microphysical process is also different in different schemes.