Abstract: Exploring the spatial correlation network of urban energy efficiency is of great significance for accelerating the green and low-carbon transition of energy, promoting regional collaborative pollution and carbon reduction, and achieving the “dual carbon” goals. The super-efficiency slacks-based measure model (Super-SBM) was used to measure the energy efficiency of cities in the Yellow River Basin, and their spatial correlations were described using the gravity model. Social network analysis and QAP regression model were applied to explore the network's characteristics and formation mechanism of energy efficiency. The results showed that the energy efficiency in the Yellow River Basin had significant spatial correlations, but the strength of connections needs to be improved. Ordos, Dongying, Yulin, Zhengzhou, and Jinan were identified as central nodes in the network, while Binzhou, Taian, Haidong, Wuzhong, and Bayannur were identified as peripheral nodes. Key "relationship hubs" in the network included links such as "Ordos→Yinchuan", "Yulin→Yinchuan", and "Yulin→Qingyang". There was a significant spillover between the four network blocks. Energy efficiency correlation network was more readily formed between cities with similar geographic proximity, population size, economic development, and level of green innovation. Additionally, average temperature, watershed proximity, urbanization rate, and technological advancement were also driving factors affecting the formation of the correlation network.