Accurate estimations of evapotranspiration (ET) are essential for understanding land-atmosphere coupling and atmosphere-underlying surface energy and water vapor exchanges. Based on input data processing, this paper simulates the temporal and spatial variation of ET in the Dajiuhu Basin from 1990 to 2018 using the BEPS-TerrainLab V2.0 model. Compared with the ET measured by an eddy covariance (EC) tower, the model explained 80.1% of the ET variation. From 1990 to 2018, the average annual ET in the Dajiuhu Basin was 1262.7 mm/yr indicating a downward trend (-27.12 mm/yr). In 2005, a sudden change point was observed based on the Mann-Kendall (MK) test and 3-year moving t-test. Around 2005, the downward trend in ET slowed and the proportional trend of ET to precipitation changed from upward trend to downward trend. Regarding spatial distribution, the ET in the basin's central part was smaller than that in the basin's surrounding area, the ET of the southern slope was higher than that of the northern slope, and the decrease in the ET rate on the sunny side was lower than that on the shady side. ET decreased as the elevation increased, with the fastest decrease observed between 2184 and 2384 m. For different landcover types, the average ET exhibited the following order: deciduous forest > mixed forest > wetland > grass > agriculture land. Decreased solar radiation is the main reason for the decreased ET in the Dajiuhu Basin, followed by increased wind speed and relative humidity, which together contribute 83.9% to the ET trend. This paper provides a theoretical basis for the study of ET changes and the mechanism of ET and provides a decision-making reference for water resource management in the Dajiuhu Basin and even the South-to-North Water Transfer Project.

No abstract