Abstract: The average annual precipitation data of 134 meteorological stations in Yunnan Province between 1993 and 2007 were treated by spatial interpolation with help of Inverse Distance Weighted method, the Ordinary Kriging method, the Anusplin Function method and Regression Model Calculation method, and then the predicted interpolation data and the observed data from the meteorological stations were comparatively analyzed. The results showed that the data obtained by all the four kinds of interpolation methods could basically reflect the spatial distribution features of the average annual precipitation of Yunnan Province. The Inverse Distance Weighted method had the minimum error and the highest interpolation precision. Although there was relatively greater error in the interpolation data, the Regression Model Calculation method could directly reflect the threedimensional characteristics of the average annual precipitation that varied with the geographical location and the altitude of the meteorological stations.