Evapotranspiration (ET) estimation is important to water resource management in the South Asia subcontinent, and remote sensing method is a good choice to get surface energy fluxes for ET estimation. However, the accuracy of regional atmospheric parameters plays a vital role especially for South Asia with few meteorological stations available. In this study, to seek a practical way to derive surface energy fluxes for ET estimation, we compare the performances of three methods based on SEBS model using different meteorological data sources (meteorological stations, GLDAS product and the combination of in situ observation, and satellite retrieval). The spatial distribution and value reasonability analysis of the estimates indicate that the three methods cannot provide reliable estimates because of the constraint of station number and GLDAS product accuracy. Correlation analysis finds that there is a good agreement between GLDAS air temperature product and in situ observations with the coefficient of determination (R2) above 0.4 for 5 days in 2008. According to the relationship, an integration method is proposed by recalibrating GLDAS air temperature with in situ observations with other atmospheric parameters supplied by GLDAS product, and the average values of the estimated LE calculated at 100 NDVI intervals are well correlated with NDVI with R-2 more than 0.695. The LE values are also within a reasonable value range with the maximum about 700 W/m(2) occurred at the full vegetation cover condition. The results indicate that the integration method is a practical way of obtaining surface energy fluxes and ET estimation for South Asia.