Enhanced understanding of environmental restoration can be achieved by examining how urban park visitors’ perceptions of biodiversity contribute to their sense of environmental rejuvenation. In this study, a questionnaire survey was conducted among 554 visitors from five representative urban parks in Fuzhou, China, and a structural equation model was used to explore the interactions among perceived biodiversity, place attachment, and self-rated repair. The main findings were as follows: (1) Perceived biodiversity had significant positive and direct effects on place dependence and self-rated restoration, but not on place identity. It is worth noting that place dependence has a deep and direct impact on place identity. (2) Self-rated restoration could be directly influenced by perceived biodiversity and place dependence. The direct impact of perceived biodiversity showed more intensity than place dependence. (3) Place dependence can be the only intermediary or link in the chain between perceived biodiversity and self-rated restoration. Conversely, place identity may not act as an independent intermediary but can play a key role in the chain of intermediaries. The study not only advances our understanding of the complex relationship between perceived biodiversity, place attachment, and self-assessed restoration; it also provides practical implications for urban green eco-design initiatives, thereby contributing to the field of urban landscape planning and formulation.

Pocket parks are small and accessible urban green spaces that may support short-term restorative experiences in high-density urban environments. However, evidence remains limited regarding how specific visual landscape characteristics are associated with psychological and physiological responses. This study quantified 11 visual landscape indicators at 30 observation points in 10 pocket parks in Fuzhou, China, using panoramic imagery and semantic segmentation. Repeated observations were collected from 50 university students during field exposure, including measures of environmental preference, perceived restorativeness, heart rate variability, and skin conductance. Multiple linear regression models were used as the primary analysis, with linear mixed-effects models conducted as a supplementary sensitivity analysis to account for the repeated-measures structure. The primary analyses indicated that higher proportions of trees, shrubs, lawns, and distant mountain views were generally associated with higher ratings of environmental preference and perceived restorativeness, whereas higher proportions of buildings, hard-paved surfaces, and fences, together with greater spatial enclosure, were generally associated with less favorable psychological evaluations. Selected physiological indicators, including heart rate, pNN50, LF/HF, and skin conductance level, were also associated with visual landscape characteristics. Stratified analyses further suggested that these associations varied across recreational space types, with vegetation-related characteristics showing particularly notable associations in architectural/structural spaces. The supplementary mixed-effects analyses indicated that some context-specific associations remained evident after accounting for repeated observations, particularly the positive associations between shrub proportion and environmental and physiological responses in architectural/structural spaces. Taken together, these findings suggest that the relationships between visual landscape characteristics and psychological and physiological responses may be context-dependent. The findings provide evidence that vegetation visibility, visual openness, and the configuration of landscape elements may inform health-supportive and restoration-oriented pocket park design. However, the observed relationships should be interpreted as short-term associations specific to the participant sample, study sites, and field conditions examined.