Evapotranspiration is one important parameter for agriculture activity, especially for water resource management. Traditionally evapotranspiration estimates by directly measured and climate data analysis. This method has low spatial representation. The utilization of the satellite image is expected to be an alternative solution to estimate evapotranspiration with a high spatial representative. This research aimed to compare Terra MODIS surface reflectance and Terra MODIS global evapotranspiration as satellite image-based evapotranspiration. This research was conducted in East Java – Indonesia. A total of 30 Terra MODIS surface reflectance and Terra MODIS global evapotranspiration image recording 2015 – 2018, MODIS geolocation that corresponding with MODIS image, topography maps on digital elevation models (DEM) format, and climate data from 3 climatological stations in East Java were used for analysis. Surface Energy Balance Algorithm for Land (SEBAL) was used to generate evapotranspiration information from Terra MODIS surface reflectance, and Penman-Monteith models used in Terra MODIS global evapotranspiration. Research shows that Terra MODIS surface reflectance and SEBAL models have better performance than Terra MODIS global evapotranspiration based on statistical methods, including mean error (ME), relative bias (RB), mean bias factor (MBF), and root means square error (RMSE).