Precipitation ensemble forecasting has great uncertainty, the uncertainty of parameters in the physical process parameterization scheme closely related to precipitation forecast is one of the sources of precipitation numerical prediction error. By introducing Stochastically Perturbed Parameterization (SPP method) on these parameters the uncertainty of representative model precipitation forecast is the frontier research field of international ensemble forecasting. In order to understand whether this method can reflect the uncertainty of numerical prediction of winter precipitation in Chinaand provide a scientific basis for business applications,based on China Meteorological Administration’s GRAPES(Global/Regional Assimilation and Prediction System)mesoscale regional ensemble prediction model,the 16 key parameters are selected from four parameterization schemes such as cumulus convection, cloud microphysics, boundary layer and near-surface layer, which have great influences on the uncertainty of model precipitation forecast,and introducing Stochastically Perturbed Parameterization (SPP),then through the ensemble prediction experiment from December 12, 2018 to January 12, 2019, a total of 31 days, comparing and analyzing the effect of SPP method on winter weather situation and precipitation ensemble prediction.The results show that the test for the random parameter perturbation method is added,the probability prediction techniques for precipitation and isobaric elements are better than control predictions without SPP method,and the improvement effect on low-level and near-surface elements is better than the improvement of the iso-surface elements in the middle or upper floors,The precipitation prediction result is superior to the control prediction test in scoring,but because it did not pass the significance test,there are no statistically significant differences.The above results indicate: under the influence of the East Asian winter monsoon,SPP method has no obvious improvement on the prediction technique of winter precipitation in China. Analyzing the reason,it may be related to the SPP method mainly represents the uncertainty of convective precipitation forecasting,but the winter precipitation process in China is mainly related to the development of baroclinic instability, model precipitation is dominated by large-scale grid precipitation, with less convective precipitation, therefore, the improvement of winter precipitation forecast is not obvious,this provides a scientific basis for applying Stochastically Perturbed Parameterization methods in the operation ensemble forecasting model.