Treatment efficacy for homeless substance abusers (primarily crack cocaine) was studied in a randomized control design with subjects (n = 176) assigned to usual care (UC) or enhance day treatment program plus abstinent contingent work therapy and housing (EC). Subjects met DSM-III-R criteria for Substance Use Disorder and McKinny Act criteria for homelessness. UC involved weekly individual and group counseling. EC involved a day treatment program consisting of daily attendance, transportation, lunch, manualized psychoeducational groups, and individual counseling. A total of 131 (74.4%) subjects (62 UC and 69 EC) were treated and followed. UC subjects attended 28.5% and EC attended 48.4% of expected treatment during the first 2 months. After 2 months, EC subjects experienced up to 4 months of abstinent contingent work therapy (44.9% of EC subjects) and housing (37.7% of EC subjects), with day treatment available two afternoons per week. Longitudinal Wei-Lachin analyses of medians (reported alcohol use, days homeless and employed) and proportions (cocaine toxicologies) were conducted across 2-, 6-, and 12-month follow-up points. EC had 36% fewer positive cocaine toxicologies at 2-months and 18% fewer at 6-months than UC with regression toward baseline at 12-months. EC had 8 days fewer days of reported alcohol use in the past 30 days, 52 fewer days homeless in the past 60 days, and 10 more days employed in the past 30 days from baseline to the 12-months. UC showed no changes except a temporary increase in employment at 6-months. This is one of the first demonstrations that homeless cocaine abusers can be retained and effectively treated.