Abstract: Climate models are a key tool for studying the mechanisms, simulation, and prediction of the East Asian winter monsoon (EAWM). The parameterization schemes for physical processes in models contain a large number of uncertain parameters, which constitutes one of the key factors limiting the ability of current climate models to accurately simulate and predict the EAWM. This study systematically investigates the contribution and mechanisms of physical parameterizations in EAWM simulations using a perturbed parameter ensemble (CAS-FGOALS-PPE) based on the new-generation climate model CAS-FGOALS-g3 developed by the Institute of Atmospheric Physics, Chinese Academy of Sciences. The results show that the CAS-FGOALS-PPE can reasonably reproduce the major climatic features of the EAWM, including key systems such as the Siberian High, Aleutian Low, northerly winds along the East Asian coast, surface temperature, East Asian trough, and westerly jet stream. Parameter sensitivity analysis reveals that parameter perturbations can substantially alter the simulated intensity of the EAWM, serving as one of the primary sources of uncertainty in its simulation. Notably, parameters such as the cloud ice-to-snow autoconversion coefficient (mg_dcs), relative humidity threshold for triggering mass flux (zmconv_rhcrit), and deep convective precipitation evaporation efficiency (zmconv_ke) significantly influence the simulation results. Deep convection processes influence the intensity of the Middle East jet stream by regulating convective activities over the northern Indian Ocean during winter, thereby exerting a significant impact on EAWM simulations. This process is governed primarily by the parameters zmconv_rhcrit and zmconv_ke. Second, microphysical processes associated with the formation of ice-phase high clouds over the Eurasian continent affect the intensity of the EAWM by modifying the land-sea thermal contrast, and these processes are regulated primarily by the parameter mg_dcs.