In urban regeneration, ecological space optimization is key for megacities to improve livability. Existing studies rely heavily on single-scale indicators such as green coverage to characterize park provision, while the multidimensional associations and spatial spillovers of park systems remain underexamined. Based on three-wave panel (2018, 2022, and 2025) data from 20 districts in Chengdu—China’s first Park City Demonstration Zone—this study constructs a four-dimensional park system index (land development intensity, supporting facility density, spatial equity, supply scale) and applies spatial Durbin models. Results show a core–periphery pattern in both park systems and livability, with park construction expanding into suburban areas faster than livability improves. Beyond these spatiotemporal patterns, the local associations between individual park dimensions and livability are multidimensional rather than uniformly positive: land development intensity and supply scale correlate positively with livability, whereas supporting facility density shows a negative association, likely reflecting, among other possible explanations, congestion-related externalities and facility-population mismatches in high-density areas. Spatial spillovers follow a “limited scale spillover, prominent configuration spillover” pattern, with spatial equity alone exhibiting significant positive cross-regional associations. This study expands park spillover research from scale-based to configuration-based perspectives and documents dimensionally differentiated park associations, providing a spatial equity viewpoint for building inclusive cities.