Under the background of in-depth application of big data and artificial intelligence, data empowerment government governance innovation has become the main way of digital government construction in various countries of the world. Under the background of government digital transformation, data sharing and opening is an important way to drive government business process reengineering, empower multiple subjects to participate in governance, and improve administrative efficiency, citizen satisfaction and government credibility. However, how to effectively adjust the path of data empowerment to improve government governance performance with government governance structure and external administrative environment is still a key issue to be discussed urgently at present. Based on resource-based theory, public value management theory, and administrative ecology theory, this paper uses panel data of 286 cities in China from 2012 to 2018 to conduct regression, mechanism and heterogeneity analysis. The results show that:(1)data resources and organizational resources significantly improve government governance performance, and organizational resources strengthen the role of data resources in improving government governance performance;(2)open government data and innovation capability significantly improve government governance performance, and innovation capability strengthens the improvement effect of open government data on government governance performance;(3)administrative ecology is the main boundary of improving governance performance of data-enabled government. In regions without administrative ecological advantages, organizational resources have a significant moderating effect, but innovation ability has no significant moderating effect. When one government promotes the improvement of government governance performance by data empowerment, it should grasp the unified relationship between data resources and organizational resources internally, and actively improve the level of open government data and the ability of multiple entities innovation externally. At the same time, it should pay special attention to its own administrative ecology and choose a suitable path to achieve accurate data empowerment. [ABSTRACT FROM AUTHOR]