Abstract: Against the backdrop of the deep integration of the digital economy into the public governance system,public data,as a strategic infrastructure for the modernization of national governance,requires effective governance,which is not only a core means to break down “data silos” and “flow barriers” but also a key support for promoting institutional innovation and the market-oriented allocation of data elements. The 2024 “Guiding Opinions of the General Office of the Communist Party of China Central Committee and the General Office of the State Council on Accelerating the Development and Utilization of Public Data Resources” clearly put forward the goal of public data circulation under the premise of security and controllability. However,in practice,the problem of governance failure caused by interest conflicts among multiple subjects is prominent:regulators face incentive imbalances,with a non-linear relationship between regulatory intensity and the willingness of suppliers to share; suppliers face a mismatch between costs and benefits,as costs such as data desensitization and storage squeeze the motivation to share; demanders face insufficient compliance constraints,leading to a negative cycle of illegal use and the shifting of governance costs. These contradictions highlight the urgency and necessity of optimizing public data governance strategies. Based on evolutionary game theory,this paper constructs a tripartite dynamic game model involving public data regulators (strategies:active regulation / passive regulation),suppliers (strategies:high sharing / low sharing),and demanders (strategies:high compliance / low compliance). Through numerical simulation using Matlab2024a,it systematically analyzes the impact mechanism of core parameters on system equilibrium. The model design strictly follows the logic of “bounded rationality-dynamic feedback-policy intervention”,incorporating key variables such as regulatory revenue (α,including incremental social benefits),regulators' own revenue (ω,such as fiscal rewards),collaborative performance bonuses (φ,additional benefits from tripartite collaboration),supply revenue (δ,direct benefits for suppliers from sharing data),supply costs (η,costs such as data cleaning and desensitization),and demand penalties (μ,intensity of punishment for illegal use) into the revenue function. It deduces the system's evolutionary trajectory through replicator dynamics equations and determines the stability of equilibrium points using the Jacobian matrix. Simulation results show that, firstly,core incentive and constraint parameters play a decisive role in the direction of equilibrium. Increasing total regulatory revenue (α+ω+φ) can cover regulators' resource input costs (γ),stimulating their motivation for active regulation; increasing supply revenue (δ) can offset suppliers' sharing costs (η),significantly enhancing the willingness for high sharing; strengthening demand penalties (μ) can make illegal costs higher than compliance costs (λ),prompting demanders to shift to high compliance behaviors. The combined effect of these three factors can drive the system to converge to a benign evolutionary stable strategy (ESS) of “high regulation-high sharing-high compliance”. Secondly,the rise in supply costs (η) has a significant inhibitory effect:when η exceeds supply revenue (δ),suppliers' enthusiasm for sharing drops sharply,and the system tends to fall into an inefficient locked state of “low sharing-low compliance”,making it difficult to realize data value. Thirdly,initial conditions have a path-dependent impact on governance effectiveness:if there is a lack of sufficient incentives (such as low ω and insufficient δ) or constraints (such as weak μ) in the early stage of reform,the system may quickly converge to a low-level equilibrium (E1(0,0,0)),and subsequent reversal would require exponential institutional costs,highlighting the importance of the policy window. The research innovation lies in breaking through the limitations of traditional static equilibrium analysis,deeply integrating abstract simulation parameters with concrete institutional tools,and constructing a closed-loop optimization framework of “parameters-tools-goals”. For regulators,it proposes to increase α+ω+φ through fiscal performance linkage (such as incorporating regulatory effectiveness into KPIs) and reputation constraints (such as inter-departmental excellence evaluations) to strengthen regulatory motivation. For suppliers,it suggests enhancing δ through benefit distribution mechanisms (such as data usage sharing) and data productization design (such as API-based services),while reducing η through common technology platforms (such as privacy computing platforms) and process reengineering (such as “one assessment,multiple reuses”) to resolve the cost-benefit mismatch. For demanders,it advocates establishing a hierarchical punishment system (such as fines,credit penalties,and data usage bans) and a compliance incentive mechanism (such as compliance whitelists and process facilitation) to stabilize μ and reduce the space for illegal activities. This study not only theoretically reveals the dynamic evolutionary laws of multi-agent strategic interactions in public data governance,addressing the core issues of interest conflicts and institutional coordination but also provides an operable path for formulating precise governance policies in practice. By dynamically adjusting parameters and matching institutional tools,it can achieve Pareto improvements in data value release and risk control,offering important theoretical support and practical references for the construction of digital governments and the market-oriented reform of data elements In response to these findings,the paper suggests a regulatory approach aiming to mitigate the adverse effects of monopolistic traffic driving,which implements interventions based on the distinctive attributes of the markets,for instance,the level of market segmentation,in which platforms operate.