Thinking regarding the needs of the chronically mentally ill has changed substantially in the post-deinstitutionalization period. Increasingly, the heterogeneity existing within this population has been recognized along with the need for an integrated continuum of residences and services. At a state level, this realization has stimulated the development of planning models which are used to predict the number and types of residential/service slots needed by the chronically mentally ill. In Colorado, a model was developed which uses data-based client profiles to predict needs along a continuum of services ranging from inpatient treatment to independent living with mental health support. In this paper, we discuss the development of the model and present the results of implementing it for a sample of 5,017 chronically mentally ill clients. The practical and heuristic value of the model is featured as well as its implications for further research.