BACKGROUND: Crime may be a significant barrier to physical activity for urban African American women, yet few studies have examined this relationship in intervention studies. This study examines relationships among neighborhood crime incidents, perceptions of crime and safety, and adherence in a walking intervention among urban, midlife African-American women. METHODS: The sample includes 148 women living in the City of Chicago. Violent crimes, disorder crimes, gun violence, and crime-related safety were examined. Adherence to walking frequency was measured as the percentage of recommended walks completed. RESULTS: Controlling for demographic characteristics and treatment group, multivariate regression analyses showed walking adherence was not associated with any of the crime measures or crime-related safety (R2 = 0.130 to 0.147). The effect of enhanced treatment did not differ by levels of objective or perceived neighborhood crime or safety. Weak to moderate bivariate correlations were observed between objective crime measures and perceived disorder crime and crime-related safety (r = 0.04 to 0.25). CONCLUSIONS: Weak correlations between perceived and objective crime measures suggest they are measuring different aspects of the crime environment. Future studies should examine perceived and objective measures in other populations and settings and other neighborhood social factors which may moderate crime and safety effects on outcomes of physical activity interventions.