Abstract: In recent years, scientific research integrity has attracted widespread attention. To promote the professionalization and normalization of research integrity management, we must first manage research data, strengthen the management of original records of scientific research work, standardize the management and audit of research data, and ensure the quality of research data. With the guidance of scientific data management “FAIR” and multi-center governance theory, the paper explores the collaborative symbiosis relationship among university research data stakeholders centered on scholars, colleges, disciplines, and schools, uses the responsibility matrix (RASCI method) to sort out the tasks, roles, division of labor, responsibilities, and so forth, then proposes a multi-center-driven university research data governance framework. The framework includes the stages for data production, architecture design, and the implementation of data governance. It covers the life cycle of research data and the life cycle of data governance. It is intended to provide a theoretical model for university research data governance.