The theoretically indicated pathways linking green space (GS) exposure to mental well-being have been consistently examined. However, few studies have deciphered the association between multiple GS exposure metrics and biopsychosocial pathways. Here, we aimed to develop a systematic exposure framework addressing different aspects of GS, including one subjective metric (self-reported GS usage) and four objective metrics, namely, quantity-based GS availability, distance-based GS accessibility, quality-based GS attractiveness, and streetscape-based GS visibility. We then explored whether, or to what extent, these various exposure metrics affect the hypothesized paths: physical activity, social cohesion, stress, and environmental stressors (i.e., air quality and noise). A cross-sectional study was conducted using a population sample (n = 1984) from Nanjing, China. Spearman rank correlation was calculated to examine the relationship between various exposure metrics. A multivariate linear regression model was used to assess the association between GS exposure and self-reported mental well-being (i.e., WHO-5 scores). Structural equation modelling (SEM) techniques were further conducted to examine the parallel/serial mediation effects of GS exposure and mental well-being. The results showed the following: (1) multiple exposure metrics were only moderately to weakly correlated, all of which were significantly associated with respondents’ self-rated mental well-being, mediated through one or more of the hypothesized pathways; (2) different exposure metrics significantly affected these mediating variables; structured GS exposure metrics were more associated with physical activity, while unstructured GS exposure metrics were more related to environmental stressors and mental stress; (3) in the parallel mediation model, stress is a relatively important mediator linking GS to mental well-being; the serial mediation model further indicated that physical activity, environmental stressors, and social cohesion worked as antecedents to reduce stress, which, in turn, was associated with better mental well-being. Our findings could help epidemiologists clarify the pathways and mechanisms by which GSs affect mental well-being, as well as assist urban planners in implementing health-based GS interventions.