<p>The evolution of urban spatial structure has profoundly reshaped the supply pattern of park green spaces. As a key pathway to enhance spatial efficiency and balance spatial justice, polycentric development has become a prevailing trend in the growth of large cities. The spatial distribution of urban populations is closely related to the accessibility and equity of park green spaces. Clarifying the coupling coordination status and influencing factors between polycentric development and park green space equity is crucial for improving residents’ well-being and promoting sustainable urban development. This study takes 102 Chinese cities as samples and establishes an analytical framework integrating exploratory spatial analysis to quantify polycentricity, the dynamic Gaussian two-step floating catchment area method to measure accessibility, and the Gini coefficient to evaluate equity. It combines a coupling coordination model with a SHAP-enhanced generalized additive model to analyze non-linear influencing factors. Results show significant hierarchical and spatial differences in park green space accessibility and polycentricity: Megacities have the optimal accessibility with high-accessibility areas distributed in a planar shape; super-large cities are divided into polarized and polycentric models; Type I and Type II large cities have relatively low accessibility. Eastern coastal urban agglomerations concentrate high accessibility, low Gini coefficients, and high polycentricity, while central and western regions exhibit low accessibility, high Gini coefficients, and low polycentricity. Coupling coordination is dominated by a moderate level, with only 3.2% of cities achieving high-level coordination. Megacities have the lowest coordination degree, and Type II large cities face the most prominent imbalance. The coupling coordination between polycentricity and accessibility is mainly characterized by imbalance and marginal coordination, with a few moderately coordinated cities concentrated in the Bohai Rim region. Among the influencing factors, the distance to the nearest park has the most significant effect, followed by the relief degree of landform and public budget, while urban area, population size, and economic indicators have weaker impacts. All key factors exhibit non-linear effects: Accessibility effectively promotes coordination within a specific range but declines rapidly below this range; a moderate increase in the distance to the nearest park can slightly improve coordination, but excessive increases lead to a sharp drop; relief degree of landform has a negative impact, and public budget shows a fluctuating upward promoting effect. The analytical framework of this study, which integrates exploratory spatial analysis, the dynamic Gaussian two-step floating catchment area method, and the generalized additive model, provides targeted support for optimizing the equitable governance of park green spaces amid urban spatial restructuring.</p>