Background: Interventions to promote walking have rarely examined how their effects varied by the attributes of the physical environment. Purpose: The purpose of this study is to examine whether perceptions of environmental walkability predicted change in walking behavior following an individual-based intervention to promote walking and whether the intervention buffered the effects of unsupportive environment for walking. Methods: Inactive adults (aged 30-65 years, 85% women) who completed a 3-month randomized control trial comparing the effect of a single mail-out of a theoretically based self-help walking program (WP, n=102); the same program plus a pedometer (WPP, n=105); and a "no-treatment" control group (C, n=107). Measures included change in self-reported walking time for all purposes and in the proportion of people reporting regular walking (i.e., ≥150 min/week and ≥5 sessions/wk). Perceptions of environmental esthetics, safety from crime, proximity to destinations, access to walking facilities, traffic, streetlights, connectivity, and hilliness were assessed at baseline and dichotomized into "low" or "high" by the median score. Covariates were social support, self-efficacy, intention to change behavior, and sociodemographic characteristics. Results: Adjusting for baseline walking, significant covariates, and study groups, walking time at follow-up was lower if streetlights or esthetics were perceived to be "low" (-24% and -22%, respectively) compared with "high" (p<0.05). In "low" esthetic conditions, those in the WPP were significantly more likely than controls to increase total walking time (Exp (b)=2.53, p<0.01) and to undertake regular walking (OR=5.85, 95% CI 2.60-12.2), whereas in esthetically pleasing environments, the between-group differences were nonsignificant. Conclusions: Walkability attributes can influence individual-based walking programs. Some environmental barriers for walking can be overcome by motivational aids.