OBJECTIVE The objective was to identify trajectories of recovery from serious mental illnesses. METHODS A total of 177 members (92 women; 85 men) of a not-for-profit integrated health plan participated in a two-year mixed-methods study of recovery (STARS, the Study of Transitions and Recovery Strategies). Diagnoses included schizophrenia, schizoaffective disorder, bipolar disorder, and affective psychosis. Data sources included self-reported standardized measures, interviewer ratings, qualitative interviews, and health plan data. Recovery was conceptualized as a latent construct, and factor analyses and factor scores were used to calculate recovery trajectories. Individuals with similar trajectories were identified through cluster analyses. RESULTS Four trajectories were identified-two stable (high and low levels of recovery) and two fluctuating (higher and lower). Few demographic or diagnostic factors differentiated clusters at baseline. Discriminant analyses for trajectories found differences in psychiatric symptoms, physical health, satisfaction with mental health clinicians, resources and strains, satisfaction with medications, and mental health service use. Those with higher scores on recovery factors had fewer psychiatric symptoms, better physical health, greater satisfaction with mental health clinicians, fewer strains and greater resources, less service use, better quality of care, and greater satisfaction with medication. Consistent predictors of trajectories included psychiatric symptoms, physical health, resources and strains, and use of psychiatric medications. CONCLUSIONS Having access to good-quality mental health care-defined as including satisfying relationships with clinicians, responsiveness to needs, satisfaction with psychiatric medications, receipt of services at needed levels, support in managing deficits in resources and strains, and care for general medical conditions-may facilitate recovery. Providing such care may improve recovery trajectories.