Abstract: It is of great significance for analyzing the spatiotemporal variation law of surface evapotranspiration and its driving factors to promote the scientific allocation of regional water resources and to do a good job of ecological water source protection. Based on MOD16 ET remote sensing data, trend analysis and significance test method were adopted to analyze the spatial-temporal variation characteristics of surface evapotranspiration in Sanjiang Plain in recent 20 years. According to Penman-Monteith formula, driving factors related to surface evapotranspiration (ET) were selected to analyze the influence of driving factors on the change of surface evapotranspiration. Ridge regression statistical model was built to analyze the main driving factors of evapotranspiration change and their relative contribution rates. The results show that the interannual fluctuation characteristics of surface evapotranspiration (ET) in the Sanjiang Plain were obvious in recent 20 years, and the overall trend of ET was increasing. ET in 91.53% of the study area shows an increasing trend, and the regional differences of ET distribution decreased year by year. The annual ET shows unimodal periodical variation, and the seasonal difference is obvious. Slope in the study area has a positive effect on ET, while elevation and wind speed have a negative effect. Temperature, sunshine duration, precipitation and NDVI are positively correlated with ET, and precipitation has the most significant correlation with ET. The determination coefficient R² of ridge regression driving analysis model is 0.823, which can effectively explain the relationship between various driving factors and ET. The results of model calculation show that precipitation and vegetation coverage have a great influence on the surface evapotranspiration in Sanjiang Plain and are the main driving forces for the change of surface evapotranspiration.