Abstract: Promoting the non-agricultural transfer of rural labor force is an important task for China to realize modernization. Improving agricultural total factor productivity plays an important role in promoting the non-agricultural transfer of rural labor force. Based on a provincial panel data from 2000 to 2020, this paper analyzed the spatial differentiation pattern of non-agricultural transfer of rural labor force and explored the impacts of agricultural total factor productivity on non-agricultural transfer of rural labor force and its spatial effects by the DEA-Malmquist model, the spatial autocorrelation method, and spatial lag model. Results show that the total factor productivity of China's agriculture shows a fluctuating upward trend from 2000 to 2020; the non-agricultural transfer of rural labor force displays an overall upward trend with significant spatial autocorrelation characteristics. The high-value agglomeration areas are mainly concentrated in the eastern coastal provinces south of the HU Huanyong line, and the low-value agglomeration areas are mainly in the western and northern border provinces. Agricultural total factor productivity can have a significant positive effect on the non-agricultural transfer of rural labor force, especially the long-distance non-agricultural transfer of rural labor force. At the same time, agricultural total factor productivity has different local and neighboring effects on the non-agricultural transfer of rural labor force, that is, promoting the non-agricultural transfer of local rural labor force and inhibiting the non-agricultural transfer of rural labor force in different places. Based on the above results, this paper suggests: increasing the input of agricultural factors, promoting the improvement of agricultural production efficiency, eliminating technical barriers, and encouraging factor exchanges among neighboring provinces.