Objective: To examine socio-demographic and psychosocial moderators, and self-efficacy as a mediator of the cross-sectional relationships between having access to recreational facilities and leisure-timephysical activity (LTPA); to investigate the extent to which the environment-LTPA associations could be explained by self-selection to neighborhoods. Design: A two-stage stratified sampling design was usedto recruit 2,650 adults (aged 20-65) from 32 urban communities varying in walkability and socioeconomic status. Participants reported perceived access to facilities and home equipment for LTPA,weekly minutes of LTPA, self-efficacy for and enjoyment of LTPA, reasons for neighborhood selection, and socio-demographic characteristics. Main Outcome Measures: Self-reported recreational walking and other forms of moderate-to-vigorous LTPA expressed in MET-minutes. Results: Specific types of recreational facilities were independently associated with LTPA. Age, education, being overweight/obese, reasons for neighborhood selection, enjoyment of, and self-efficacy for LTPA moderated these relationships. Self-efficacy was not a significant mediator of these cross-sectional associations. Conclusion: These findings have potentially significant implications for the planning of environmental interventions aimed at increasing population-level LTPA particularly in those who are less attitudinallyinclined to being physically active.