Recent advances in remote sensing technology have led to ways of estimating regional evapotranspiration from AVHRR satellite images. The newly developed MODIS satellite is being increasingly used in estimating evapotranspiration. This study aims to estimate the meteorological characteristics of Penman-Monteith method from AVHRR and MODIS images and then the evapotranspiration results are compared. Four major meteorological characteristics (air temperature, vapor pressure deficit, wind speed, and net radiation) were derived from images and station observations. The average errors of the evapotranspiration estimation were 1.197 mm/day (AVHRR) and 0.63 mm/day (MODIS), and the relative root-mean-squared errors were 0.477 (AVHRR) and 0.288 (MODIS). The verification results revealed that both MODIS and AVHRR images can be applied to estimate evapotranspiration, in which the estimation using MODIS images is better. Amount these four meteorological characteristics, the net radiation has the best estimation and the win speed was the poorest. However, the most sensitivity of the net radiation has led the worst result in estimating evapotranspiration. The accurately estimation of the net radiation is necessary in the future. Totally, the influence by the error of each meteorological characteristic was less than 11%. This suggested that this method with MODIS images is useful and practical in regional potential evapotranspiration estimation and can be suitably used in ungauged areas.