Evapotranspiration (ET) is an important geo-biophysical parameter for water management in watersheds. The determination of ET generally involves the use of ground-based meteorological data, which does not adequate capture the spatial patterns of mass and energy fluxes, especially in mountainous areas. In this work we evaluate the daily spatial distribution of ET over mountainous watershed in southeastern Brazil, by coupling Surface Energy Balance Algorithms for Land (SEBAL), global solar radiation (GSR) model and a gridded weather dataset. We used OLI/Landsat-8 surface reflectance and TIRS thermal images, SRTMGL1 Digital Elevation Model and weather gridded data from Global Land Data Assimilation System (GLDAS). To estimate daily tilted GSR, an analytical parameterization was applied using the relation between terrain and sun angles over 24h integration time. Daily ET was estimated through SEBAL model, adapted for tilted surfaces, based on the estimated GSR and resampled GLDAS data. Tests were performed in summer/wet (01/12/2015) and winter/dry (09/25/2015) periods to evaluate the seasonal differences in ET over tilted surfaces. The results indicated different spatial patterns of daily ET on the watershed in each period. In summer, ET was 9.8% higher on slopes facing the South while in winter, ET was 10.6% higher on slopes facing North and East. High variability in daily ET was found on steeper slopes (above 45°), with mean coefficient of variation (CV) of 25% and 43.6% for summer and winter, respectively.The notable spatial heterogeneity of daily ET over this watershed indicate the complex partitioning of mass and energy fluxes from different terrain angles, which can influence hydro-ecological processes at local scale. The presented approach allowed a more detailed capture of the spatial variability of ET in a complex terrain, which can be useful in mountainous watersheds with scarcity ground-based data.