Mega-cities, while leveraging large populations for development, need to rationally plan spatial resources, especially for optimizing elderly care services. This study uses Chengdu, Sichuan Province, introducing factors such as accessibility of elderly-friendly medical institutions, weighted built environment factors, and selection influence factors considering service capacity levels and residents' maximum travel time into an improved potential model to explore the accessibility of elderly care services. The study found that the outskirts of Chengdu have a higher aging population, while the central city and its surrounding areas have a larger older population but lower aging rates. Elderly care facilities are mainly small and medium-sized, exhibiting a radial layout centered around the main urban area. The overall level of spatial accessibility to elderly care service resources is relatively low. The central city and its surrounding areas form high-accessibility zones, while the remote suburbs have large areas of low accessibility and scattered high-accessibility points. Based on the findings, four areas for future urban planning and resource allocation were identified: development retention areas, urgent optimization areas, configuration compensation areas, and remote special areas. This study provides new insights into the classification of elderly care resource accessibility and fairness evaluation, offering a reference for planning studies in other mega-cities.