The middle reaches of Yangtze River are full of agricultural resources and have been contributing greatly to China's food security. However, in recent years, the green development of agriculture in the middle reaches of Yangtze River has faced great challenges due to the promotion of agricultural modernization. Meanwhile, with the advancement of regional integration, the cross-regional flow of production factors such as technology, capital and labor in the middle reaches of Yangtze River made the inter-regional agricultural economy also present a complex network relationship. Therefore, it is necessary to explore the spatial correlations of agricultural green development in the middle reaches of Yangtze River from a network perspective, which is significant for promoting the green coordinated development of agriculture. In this study, to evaluate the level of agricultural green development in the middle reaches of Yangtze River, the super-efficiency MINDS model was used to measure the agricultural eco-efficiency from 2001 to 2019. Moreover, the gravity model was used to identify the spatial correlation of agricultural eco-efficiency, and the social network analysis method and QAP method were used to reveal its network structure characteristics and driving factors. The results showed that: (1) The spatial correlation of agricultural eco-efficiency in the middle reaches of the Yangtze River showed a complex network structure. The spatial correlation network had high accessibility and the hierarchical structure tended to be loose, but the network stability tended to decline. (2) Wuhan and Changsha occupied the core position in the spatial correlation network, acting as both central actors and as ″intermediaries″ and ″bridges″; the provincial border cities, such as Yueyang, Huangshi, played the role of marginal actors in the network. (3) The core-periphery structure of the spatial correlation network evolved from large aggregation and small dispersion in the Tenth Five-Year Plan period to large dispersion and small aggregation in the Thirteenth Five-Year Plan period, and the core area changed from aggregation in the middle to divergence around. (4) The adjacency of geographical space, the similarity of the status of regional agricultural industry, fiscal expenditure, economic development and the difference of transportation all contributed to the formation of spatial correlation network.