The reference evapotranspiration (ET0) is an essential variable in the agrohydrological systems and its estimation on a regional scale is limited to its spatial variability. This study compares two approaches for preparation of spatial distribution maps of ET0 in Mazandaran province of Iran. In the first approach, ET0 was calculated using climatic data and Hargreaves-Samani equation in weather stations locations and then were interpolated. In the second approach, the components of the Hargreaves-Samani equation were interpolated and then ET0 maps were prepared by applying the Hargreaves-Samani equation and suitable commands in GIS. The 10-year climatic data for 51 weather stations (46 stations for preparing ET0 maps and 5 stations as validation station) were gathered over Mazandaran province. Semivariograms were calculated and the best semivariogram model was selected on the basis of the least value of Residual Sums of Squares (RSS). The spatial correlation of the data was compared on the basis of Nugget to Sill ratio. The data were interpolated using Ordinary Kriging method and the interpolation error was computed by cross validation technique based on Root Mean Square Standardized Error (RMSSE). The predicted ET0 values were compared to the computed ET0 in validation stations and sensitivity analysis was conducted. Results show the second approach had better spatial correlation and lower interpolation error and the difference between these two approaches were not significant. Therefore, the accuracy of the ET0 maps is more related to the method of computing ET0 than the type of climatic data is being interpolated.