In the context of the "Park City" construction and refined governance in Shanghai, there is still a need to explore the evaluation methods and governance strategies for the service efficiency of urban green spaces, merging subjective and objective aspects. This study introduces theories and methods related to landsenses evaluation to develop a landsenses evaluation system for urban green spaces that connects objective service efficiency indicators with subjective perception experience indexes, identifies sensory space units, and suggests governance strategies. The research focuses on Yangpu District in Shanghai, obtaining objective evaluation results of green space service efficiency through a GIS platform, alongside subjective evaluation results derived from crawling and sentiment analysis of online comment text data. By spatially overlaying subjective and objective evaluation results and performing spatial clustering and autocorrelation analysis, four types of landsenses units of urban green spaces are identified based on the comprehensive landsenses evaluation. The findings indicate that the overall landsenses evaluation of park green spaces in Yangpu District is relatively high; however, ribbon-like lowlands exist in the central and southern regions, displaying significant spatial aggregation characteristics in the spatial pattern of the coupled subjective-objective evaluation. This underscores that existing park city governance strategies have seen initial success, yet targeted sensory optimization is necessary for different landsenses units of urban green spaces. The study provides recommendations for optimizing urban green space planning and enhancing citizens' experiences of green space perception.