More and more attention has been paid to the influence of the uncertainty of the physical processes on the prediction accuracy. In this paper, multivariate analysis of variance and Tukey"s test are used to quantify the model uncertainty related to the parameterization of physical processes and to select the optimal parameterization scheme combination. Taking a rainstorm in South China from May 21 to 22, 2020 as an example, the microphysics scheme, boundary layer scheme and cumulus convective scheme, which are closely related to precipitation, are selected to carry out 80 sets of experiments. The results of multivariate analysis of variance show that the microphysics scheme is the most important for precipitation simulation results and model prediction results, and the interactions between different physical processes cannot be ignored. Tukey"s test is used to compare the discrepancy among different parameterization schemes, and the optimal parameterization scheme combination is selected, which is WSM7 scheme + YSU scheme + GF scheme. In addition, this paper compares the differences of different microphysics schemes, and analyse the influence of the differences of microphysics schemes on the precipitation simulation error from the perspective of hydrometeors and microphysics conversion processes. The results show that hail plays an important role in the simulation of the extreme precipitation. The change of rainwater mixing ratio is mainly due to the melting term of ice particles. The evaporation of rain affects the intensity of cold pool through latent heat absorption, and then affects the subsequent precipitation propagation, finally affects the distribution of heavy precipitation.