Abstract: The influence of land surface schemes on precipitation simulation is examined through two multiyear simulations of the China's climate during the period from 2001 to 2005. The Regional Climate Model (RegCM Version 4.0) coupled to the Biosphere-Atmosphere Transfer Scheme (BATS) and the Community Land Model (CLM) respectively is used for the climate simulation. The precipitation simulations are compared to two precipitation datasets Climate Research Unit of University of East Anglia (CRU), and Global Precipitation Climatology Project (GPCP). The result shows that both spatial distribution and temporal variation of China's precipitation can be reasonably simulated by RegCM4.0 coupled to the two land surface schemes. However, the RegCM_BATS generally produces a positive bias in the simulated precipitation, while the opposite is true for the RegCM_CLM precipitation, indicating that precipitation simulation is sensitive to different land surface schemes. The largest positive bias is found in Northeast China for RegCM_BATS, and in the South of Yangtze River for RegCM_CLM, the RegCM_CLM precipitation has the biggest negative bias. The simulation bias is large in summer months, and small in winters. The precipitation differences between the two simulations decrease from Southeast to Northwest. Both RegCM_BATS and RegCM_CLM can well simulate the precipitation frequency for different intensities. RegCM_CLM generally overestimates low intensity precipitation and underestimates high intensity precipitation. The opposite is true for RegCM_BATS. Differences in the modeled evaporation and latent heat flux with the two land surface schemes are the major causes of the precipitation difference between the two simulations. The influence of evaporation on precipitation simulation is larger during the summer months than that of winter months. While, the water vapor advection has an insignificant influence on precipitation simulation.