Park green spaces serve irreplaceable functions in urban green infrastructure systems by sustaining ecological equilibrium, enhancing landscape configurations, and elevating resident living standards. This research investigates park green spaces within elderly-friendly community life circles in Hangzhou's Shangcheng District. Integration of multi-source datasets-including remote sensing imagery, mobile signaling data, POI (points of interest), and urban road networks-enabled construction of a comprehensive geodatabase. Application of kernel density estimation, buffer zone analysis, and network analysis, combined with entropy-weighted TOPSIS for weight calculation, the study systematically evaluates variations in urban park green space service levels across different scales and explores their underlying causes. The findings aim to provide a scientific basis for improving green space service levels. The main conclusions are as follows: (1) The overall park green space service level of Shangcheng District at the regional scale is at level II. Park green spaces demonstrate reasonable structural composition, balanced spatial distribution, appropriate scaling, commendable attractiveness, and environmentally friendly attributes. However, there are also problems of unbalanced distribution of green space resources. Targeted improvements and focused attention are imperative to ensure equitable distribution and efficient utilization of green space resources. (2) At the street scale, disparities in park green space service levels are observed. Qingbo and Dinglan streets exhibit high service levels (Grade Ⅰ), offering elderly residents diverse leisure spaces and social venues. In contrast, Jianqiao, Pengbu, and Jiubao streets show low levels (Grade Ⅴ). Service levels in other streets fall between Grades Ⅰ and Ⅴ. Urgent improvements are required to enhance service quality and promote balanced development of green space resources across all areas. (3) At the community scale, the distribution of park green spaces shows a trend of polarization and clustering. 14.9% of parks achieve high service levels (Grade I), concentrated around West Lake, along the Qiantang River, and in northern Shangcheng District, while sporadic high-level green spaces in Garden Community, Shuixiang Community, and Zhangjiaba Community are scattered in central areas. Conversely, 19.8% of parks demonstrate low service levels (Grade V), predominantly clustered in central and northeastern regions. The research constructs innovative assessment dimensions that deeply integrate elderly-specific evaluation parameters, precisely identifies service gaps within targeted population groups, and provides theoretical and methodological support for follow-up work such as age-friendly urban planning and aging-adaptive facility retrofitting. Future urban park green space planning and construction should fully consider the behavioral characteristics of the elderly as a foundation, construct all-age-friendly circular vitality networks, improve the spatiotemporal accessibility and spatial utilization efficiency of green spaces, and ensure that the elderly population can equally enjoy the ecological benefits of community life circles.