The aim of this research was estimating the accuracy of CropWat software to calculating potential evapotranspiration (ET0) in arid and semi-arid region of Iran. For this purpose, 9 locations of arid and semi-arid regions of Iran selected and the grass lysimeter data collected too. The lysimeter data are collected duration of May through November during 1988-1997. For the comparisons of the ET0 lysimeter data and ET0 resulted CropWat software in different point, three statistical parameters were used include Mean Absolute Relative Error (MARE), Root Mean Square Difference (RMSD) and correlation coefficient (R2). The results of this research show that the average of MARE, RMSD and R2 computed about 31%, 2.3 and 0.7, respectively in the study areas. In addition, the results of this research show that when the average of wind speed was less than 1 m sec-1 or wind speed was more than 1 m sec-1 with low ET0 lysimeter (ET0 less than 6 mm per day), CropWat has a low sensitive for estimating ET0 and it's necessary to improve the results for these areas. In concluded, the results of this research were showed that in these area because of intensive temperature and solar radiation, CropWat can not be estimated ET0 exactly.