Assertive community treatment (ACT) programs originated to meet the needs of the severely mentally ill (SMI) for multiple psychosocial supports necessary for community living. Housing has been prominent among these needs and research to date on ACT programs has addressed residential outcomes, principally reduced homelessness and maintenance in some form of stable housing. However, limited studies have examined variables that might predict particular types of residential placements, which range from unsupervised to highly structured environments. The purpose of this study was to identify factors related to residential placement for the subjects in a model ACT program. Of multiple variables assessed, two were significantly related to residential placement. Chronic respiratory illness was strongly associated with placement in supervised group homes. Team assignment was associated with placement, with clients in one of three teams significantly more likely to live more independently. Clients assigned to this team showed higher psychosocial functioning than others in terms of functional ratings on the Multnomah Community Ability Scale. However, higher function alone was not significantly associated with placement, suggesting a role for clinical judgement and a need for future research in this area.