OBJECTIVES: Women in impoverished inner-city neighborhoods are at high risk for contracting HIV. A randomized, multisite community-level HIV prevention trial was undertaken with women living in 18 low-income housing developments in 5 US cities. METHODS: Baseline and 12-month follow-up population risk characteristics were assessed by surveying 690 women at both time points. In the 9 intervention condition housing developments, a community-level intervention was undertaken that included HIV risk reduction workshops and community HIV prevention events implemented by women who were popular opinion leaders among their peers. RESULTS: The proportion of women in the intervention developments who had any unprotected intercourse in the past 2 months declined from 50% to 37.6%, and the percentage of women's acts of intercourse protected by condoms increased from 30.2% to 47.2%. Among women exposed to intervention activities, the mean frequency of unprotected acts of intercourse in the past 2 months tended to be lower at follow-up (mean = 4.0) than at baseline (mean = 6.0). These changes were corroborated by changes in other risk indicators. CONCLUSIONS: Community-level interventions that involve and engage women in neighborhood-based HIV prevention activities can bring about reductions in high-risk sexual behaviors.