Abstract: In order to comprehensively reflect the characteristics of space evolution and its formation mechanism of agricultural energy consumption in China from the view of temporal and spatial combination, according to agricultural energy consumption per cultivated area in all provinces, using the index of global spatial autocorrelation and the index of local spatial autocorrelation, the study analyzed the spatial distribution pattern of agricultural energy consumption in China and established the spatial regression model to analyze the influencing factors. The results showed that the level of agricultural energy consumption in China was increasing, but the inter-provincial absolute and relative difference was expanding; there was positive spatial autocorrelation in agricultural energy consumption which was demonstrated by the agglomeration of high-value and low-value regions. The cold-spots were mainly located in the vast Midwest region and the hot-spots were mainly located in southeast coastal areas, especially in Yangtze River Delta region, while the warm-spots were located in the region that bordering the east, the middle and the west. Agricultural workers, total power of agricultural machinery, agricultural output and agricultural investment in fixed assets were the main causes of spatial differentiation of agricultural energy consumption. But the impact of various factors was different. In addition, the spatial autocorrelation was also an important factor that affected the differentiation of agricultural energy consumption.