Abstract: To quantitatively study the component-evapotranspiration in the process of surface water conversion in Beijing, we built the SEBAL model through ERDAS based on Landsat7 ETM + remote sensing image information combined with meteorological data, calculated that the average evapotranspiration on July 26, 2014 in Beijing was 2.49 mm, and verified the reliability of the SEBAL inversion results by using the Penman formula and MOD16 product data. The results showed that: the error of evapotranspiration calculated by SEBAL model and Penman’s formula method and MOD16 data product was 3.2% and 3.3%, respectively, which had certain reliability, and the data inversion by Landsat7 had the advantages of convenient data acquisition and high spatial resolution; by comparing the evapotranspiration distribution of each region, it was concluded that the overall evapotranspiration in Beijing was high in the northwest mountainous area and low in the southeast plain which resulted from the high coverage of forest vegetation in the northwest mountainous area; through statistical analysis, the evapotranspiration of the water area was the highest, and that of the hardening area and the building were close to zero; after removing the water body and the hardening area, the