The high spatial-temporal resolution of satellite precipitation data can be widely used in flood disaster monitoring and hydrological modeling. The accuracy assessment and reliability analysis of the satellite precipitation data are still the important research issues. The paper overall evaluated the performance of TRMM satellite precipitation product 3B42V7 in Hunan Province by using the gauged precipitation data from 23 meteorological stations as the benchmark data. From the different time scales of daily, monthly, seasonal and annual,and different space elements such as spatial distribution and elevation,the paper compared and analyzed adaptation situation of satellite rainfall products from 1999 to 2012 in Hunan Province.The results show that the performance of satellite precipitation data on daily scale is not satisfactory, the correlation coefficient is only 0.31. But it goes to be much better on monthly scale with the correlation coefficient of 0.88. the annual rainfall from TRMM 3B42 can correlate well with the amount of rainfall on the ground of the region. The performance of satellite precipitation data in arid season (from November to February) is better than that in humid season (from May to August).Besides, the satellite precipitation data of 3B42 exists obvious spatial variability. The spatial factors such as elevation and spatial distribution have greater influence on the accuracy of satellite precipitation data than that of the amount of rainfall.