Abstract: A self-developed moist physics parameterization package (PhysCN) is implemented in the Global-to-Regional Integrated forecast SysTem (GRIST) model. PhysCN consists of convective parameterization, macrophysics condensation and microphysical processes. Impact of the moist physical processes on climate modeling are evaluated by comparing with the default schemes (PhysC) based on the 10-year Atmospheric Model Intercomparison Project (AMIP) type simulation. Results show that PhysCN can reproduce mean tropical precipitation and its seasonal variation more realistically. Compared with PhysC, PhysCN reduces artificial precipitation over the intertropical convergence zone (ITCZ) regime, the tropical Pacific and the Indian Ocean. It also decreases the double ITCZ bias. There exist more ice clouds near the top of the tropical troposphere and the mid- and high latitudes in the PhysCN simulation than in the PhysC simulation, which enhances longwave cloud radiative forcing. However, shortwave cloud radiative forcing is weakened in the PhysCN simulation, which increases the bias of global net radiation budget. This is because the single ice microphysics approach uses a single prognostic category to represent cloud ice and snow mixing ratio, while the cloud macrophysics scheme diagnoses ice cloud fraction based on total relative humidity over ice particles. Low clouds increase and become denser, especially in the mid- and high latitudes. This study exhibits the stability and reasonableness of this moist physics package in GRIST, and shows that interaction between the microphysics and diagnosed cloud needs to be improved.