A longitudinal experimental design was used to compare the effectiveness of three community-based treatment programs serving homeless mentally ill people: traditional outpatient treatment offered by a mental health clinic, a daytime drop-in center, and a continuous treatment team program that included assertive outreach, a high staff-to-client ratio, and intensive case management. At 12-month follow-up, clients in all three treatment programs spent fewer days per month homeless, showed fewer psychiatric symptoms, and had increased income, interpersonal adjustment, and self-esteem. Clients in the continuous treatment program had more contact with their treatment program, were more satisfied with their program, spent fewer days homeless, and used more community services and resources than clients in the other two programs.