Abstract: Green transformation of farmland use can be expected to improve grain production efficiency at the regional scale. It is therefore necessary to investigate the spatial correlation network of the green transformation of farmland use and its impact on grain production efficiency. However, existing studies have rarely examined the network externalities of spatial correlations in the green transformation of farmland use, particularly from the perspective of network node attributes. This study aims to explore the influence of node status and role positioning within the network on grain production efficiency. The spatial correlation network was first determined for the green transformation of farmland use. Network structure was then formed to improve grain production efficiency. A case study was conducted in Jiangsu Province, China. Empirical tests were conducted using the super-efficiency slacks-based measure (SBM) model, modified gravity model, social network analysis, and panel regression. The results show that: 1) A positive trend was observed in the green transformation of farmland use and grain production efficiency at the county scale from 2006 to 2022. Specifically, 19 counties reached the medium-high or high level for the green transformation of farmland use by 2022, and more than half of the counties achieved medium-high or high grain production efficiency. Nevertheless, spatial imbalance was required to improve after evaluation. 2) The spatial correlation network shared the multi-threaded interweaving core-periphery structure. Network density remained between 0.20 and 0.25, indicating a first increasing and then decreasing trend, whereas network efficiency showed the opposite pattern. Although inter-county spatial linkages in the green transformation of farmland use have gradually emerged, these linkages still need to be further strengthened. 3) Inter-county differences in individual network structure were observed in degree centrality, closeness centrality, and betweenness centrality, with several counties exhibiting substantially higher node centrality than others. Furthermore, Sheyang County, Yancheng Municipal District, and Gaoyou City also shared high degree centrality and betweenness centrality, indicating the function of leaders and intermediaries within the network. Counties with high closeness centrality were concentrated in southeastern Jiangsu, indicating stronger accessibility and rapidly greater connection. 4) Degree centrality, closeness centrality, and betweenness centrality all exerted significantly positive effects on grain production efficiency. Specifically, counties with high degree centrality reduced input redundancy in grain production after network agglomeration; those with high closeness centrality improved resource allocation using strong linkages; and those with high betweenness centrality enhanced the matching of grain production resource supply and demand after network regulation. Together, these network functions contributed to higher grain production efficiency. The peer effects of structural optimization, resource intensification, and technological innovation were identified as key mediating channels. Moreover, the network degree centrality exerted a stronger positive effect on grain production efficiency in counties implementing holistic promotion of grain yield improvement. Greater efforts should be made to promote regional cooperation in the green transformation of farmland use by leveraging network externalities to improve grain production efficiency. These findings can provide decision-making references for farmland protection and national food security.