We investigated the relationship between walking levels and the local neighbourhood physical environment during the Walking for Wellbeing in the West (WWW) randomised pedometer-based community intervention. Walking activity was recorded as step counts at baseline (n = 76), and at 3 months (n = 57), 6 months (n = 54), and 12 months (n = 45) post-intervention. Objective physical environment data were obtained from GIS datasets and street surveys conducted using the SWAT audit tool. Sixty-nine environment variables were reduced to eight environment factors using principal axis factoring, and the relationship between environment factors and (i) step counts, and (ii) the change in step counts relative to baseline, was examined using hierarchical multiple linear regression, controlling for age, gender, income, and deprivation. Five environment factors were significant predictors of step counts, but none were significant predictors of the change in step counts relative to baseline. None of the demographic variables included in the analysis were significant predictors at any stage of the study. Total variance explained by the environment ranged from 6% (P < 0.05) to 34% (P < 0.01), with lowest levels during the initial stages of the study. The physical environment appears to have influenced walking levels during the WWW intervention, and to have contributed to the maintenance of walking levels post-intervention.