How to achieve high-quality government data openness and fully leverage the value of data elements have emerged as a critical concern for both academia and industry in digital age. Based on panel data from 21 provinces, this study introduces the dynamic capability theory to establish a “capability-performance” analytical framework. By employing a mixed-method approach that integrates fuzzy-set qualitative comparative analysis (fsQCA) and regression analysis, this study explored the separate impact of each capability and their interactive relationships on government data openness performance. The results shows: capabilities including institutional design, technology assimilation and data provisioning have a significant positive impact on government data openness performance; these capabilities cannot independently constitute the necessary conditions for the high-performance of government data openness; these capabilities form three conditional configurations leading to the high-performance of government data openness : the “opportunity perception” type, the“comprehensive driving” type and the “rapid response” type.