Data sharing serves as a crucial link for the integration of data resources among governments, platforms, and enterprises, with data granularity playing a key role in optimizing the efficiency of sharing. This paper constructs a tripartite evolutionary game model involving platforms, enterprises, and governments to explore the impact of data granularity and other factors on the sharing strategies of each entity. The research indicates that reducing data granularity can enhance the initial willingness of platforms and enterprises to share. When data granularity is below a certain threshold, the system tends towards an“ideal sharing” equilibrium. If data granularity meets the critical conditions and regulatory costs are low, the government is inclined to adopt an “incentive first, then regulation” empowerment model to achieve the “ideal” equilibrium. Conversely, if regulatory costs are high, the government prefers a “regulation first, then incentive” model. Further research shows that increasing government incentives, reward coefficients, and platform subsidies can accelerate the system’s evolution towards the ideal equilibrium. The research conclusions provide a solid theoretical support and decision-making reference for promoting data sharing practices in a data circulation environment.