Because that the terrain is complex and the station is scarce in the Heihe River Basin, understanding the fine spatial distribution and the amount changes of precipitation become a difficult problem. This paper focuses the influence of the atmospheric variability in the earth-atmosphere system on the precipitation to solve the above problem. Firstly, The paper designed a 16-member sensitivity test by a cross pairing of different parameterization schemes of land surface process, planetary boundary layer and microphysics. Based on which, we simulated two representative precipitation events in the Heihe river basin, and verified the simulation effect with the daily precipitation data observed by 12 meteorological stations in this area. The sensitivity tests showed that the WRF model have a good performance to the topographic effect of precipitation process. Secondly, using the best combination of parameterization schemes, we get the data of the daily precipitation and make the precipitation as a function of latitude, longitude, slope. Combined the model output with DEM analysis of the daily precipitation, We present a method to calculate the grid distribution of precipitation(100m×100m) and the corresponding results, by "dynamical - Statistics" downscaling methods. This paper gives a different approaches and has a reference value to the research of the fine spatial distribution in the Heihe River Basin.