The excessive use of inputs per unit of agricultural land poses a great threat to ecological sustainability. Using an input‐oriented data envelopment analysis (DEA) model, this study analyzes ecoefficiency of intensive agricultural production in 31 provinces in China. The results show that the total efficiency of only six provinces can be considered fully efficient and that scale efficiencies are generally lower than technical efficiencies. Then, the spatial distribution of ecoefficiency is analyzed. The findings demonstrate that the provinces whose ecoefficiencies are maximal are primarily located in western China. The technical efficiencies in the western region are better than those in the eastern and middle regions. Imperfect scale efficiencies are distributed across all three regions. Furthermore, using the Tobit model, an analysis of the factors that influence ecoefficiency shows that the variables of farmland area per capita (FA), income per capita (IC), population per household (PH), and population burden coefficient (PB) have statistically significant impacts on total efficiency. The distinct effects of the variables on total efficiency are caused by their differential effects on technical efficiency and scale efficiency. Finally, suitable policies designed to improve ecoefficiency are proposed according to the local circumstances of each of the three regions.