Objective: The study evaluated the success of a generic service program f or older homeless persons in improving the well-being of mentally ill clients. It attempted to identify factors that predicted the number of service encounters and outcome in seven areas, including housing, entitlements, physical and mental health, and sobriety. Methods: All persons newly admitted to the program during a two-and-a-half-year period were asked to participate in structured interviews assessing their physical and mental health and their support networks. Two-thirds of those eligible, or 130, participated in the intake interviews. At three-year follow-up or last contact, outcome was compared for 41 psychiatric clients with psychotic symptoms or self-reported history of psychiatric hospitalizations and 89 clients with no psychiatnic symptoms or previous hospitalizations. Results: Persons with mental illness averaged 2.5 favorable outcomes, and a majority obtained temporary or permanent housing, improved their physical health, and secured entitlements. However, mentally ill clients bad significantly fewer service encounters and favorable outcomes than clients who were not mentally ill. Types of presenting problems at intake were the only significant predictors of outcome for mentally ill clients. Conclusions: A generic service program for older homeless persons can successfully improve the well-being of mentally ill clients, although outcomes are less favorable for such clients than for clients who are not mentally ill.