This paper compares five spatial interpolation methods of Kriging-exponential, Krigingspherical,Kriging-Gaussian with different weighting values are unity and 1/Standard Deviation in temperature interpolation,utilizing GIS software spatial analysis functions. The research data are the minimum temperature, mean air temperatures and maximum temperature of the first ten days of January, April, July and October from 1990 to 2005 in 14 weather stations of Alaska. The result shows that Kriging-doublespherical,Matern and Gaussian interpolation methods are the highest-accuracy methods for unity weightage, Matern and Gaussian interpolation methods are highest-accuracy methods for 1/Standard Deviation weightage and Matern with unity and 1/variance weightage is best suitable interpolation methods for the spatial data