Abstract:High precision spatial information of the temperature is very important for utilizing the regional energy resource and adjusting the agricultural structure. Based on technology platform of geography information system (GIS) and geo-statistics technology, mean annual temperature of 72 meteorological stations from 1971 to 2008 has been used as the data source, the following methods, such as inverse distance weighing interpolation, ordinary kriging interpolation, spline interpolation and multiple regression interpolation were used to spatially interpolate mean annual temperature in study area. Interpolation results were analyzed by the prediction accuracy verified by 10 validated stations. The result demonstrated that mean relative error, mean absolute error, root mean square error of multiple regression interpolation were respectively 0010,0173 and 0221,its error was less than other three interpolation methods, and the interpolation result of multiple regression interpolation was the best of all. The simulating value of mean annual temperature of validated stations was close to the actual value.