Background: The existing literature predominantly concentrates on the street or town level, often neglecting the resources embedded within communities.Methods: This study investigates multiple tiers of districts, streets (towns), and housing estates, while integrating community-based beds. We utilized several spatial analysis techniques: buffer analysis, grouping analysis, create fishnet tool, and the genetic algorithm. Data management and analysis were performed using ArcGIS 10.8.Results: A significant disparity in the spatial allocation between pension resources and the ageing population is evident in downtown Shanghai as of 2020. By establishing a 15-minute community life circle, this research identifies areas with inadequate coverage of pension resources.Conclusions: This study examines the site optimisation of pension facilities using a dynamically updated genetic algorithm. For comparison, three other methods were also utilized: fishnet label point, grouping analysis, and the standard genetic algorithm. The results reveal that the dynamically updated genetic algorithm emerged as the most effective approach, achieving full coverage of the 15-minute community life circle with the fewest new pension facilities. This research aims to provide recommendations and insights for the overall urban planning and pension resource allocation in Shanghai.