The spatial and temporal distribution of agricultural gray water footprint efficiency of 35 prefecture-level cities in Huai River Basin was studied by using the gray water footprint efficiency calculating model. Kaya identity and LMDI model were employed to decompose the driving factors of agricultural grey water footprint efficiency into five aspects: agricultural economic effect, fertilizer intensity effect, gray water output scale effect, agricultural environmental effect, and cultivated land resource effect. A classification study of driving models of agricultural gray water footprint efficiency in every city in Huai River Basin was conducted. The results show that the agricultural gray water footprint efficiency of 35 prefecture-level cities in Huai River Basin has been significantly improved during 2000-2015. According to the distribution of each province, the agricultural gray water footprint efficiency of Shandong Province is the highest, followed by Anhui Province and Jiangsu Province, with Henan Province being the lowest. Among the driving factors of agricultural gray water footprint efficiency, agricultural economic effect, gray water output scale effect and cultivated land resource effect are positive, while chemical fertilizer intensity effect and agricultural environment effect are negative. Among them, agricultural economic effect is the most critical driving factor to improve the efficiency of agricultural gray water footprint. The driving models of agricultural gray water footprint efficiency of 35 prefecture-level cities in Huai River Basin can be divided into 4 categories: single-factor driving model I dominated by agricultural economic effect, single-factor driving model II dominated by gray water scale output effect, two-factor driving model III and three-factor driving model IV. Different governance measures should be taken in regions with different driving models.