[Objective] The spatiotemporal differentiation and evolution characteristics of the green efficiency of agricultural water use in China and the current status of Chinese agricultural development were analyzed in order to further promote high-quality development of agriculture. [Methods] The super-efficiency evidence-based model (EBM) was used to estimate the green efficiency of agricultural water in 31 provinces and cities during 2001-2020. On this basis, kernel density estimation (KDE) was used to conduct non-parametric tests, and ArcGIS graphics were used to explore the spatiotemporal differentiation and evolution characteristics of the green efficiency of agricultural water use in different regions of China. Finally, the geographical detector method was used to investigate the comprehensive influence of different driving factors on the green efficiency of agricultural water use. [Results] ① From 2001 to 2020, the national green efficiency of agricultural water use (the annual average of each province) initially increased and then decreased. ② The national green efficiency of agricultural water use showed a spatial pattern of "lower in the north and higher in the south", "lower in the west and higher in the east". ③ The regional gap of the green efficiency of water use in China showed a narrowing trend, with dynamic convergence characteristics. The green efficiency of agricultural water use in each province was polarized to varying degrees, among which the nuclear density curve in Northwest China was the most gentle, and the polarization was the most serious. ④ Technical level, resource endowment, and ecological environment had greater influence on the green efficiency of agricultural water use than economic development level. [Conclusion] Each province should combine its own advantages and promote the green efficiency of agricultural water use from the aspects of agricultural technology level and rural social welfare.