Few studies have examined physical activity in patients who are severely mentally ill. We assessed feasibility and validity of using standardized self-report and objective measures of physical activity in 20 community-dwelling veteran patients, most with psychotic disorders and substance abuse in remission. Test-retest reliability of a past week physical activity survey and the Community Healthy Activities Model Program for Seniors interview were evaluated with repeated interviews. RT3 accelerometer data provided objective physical activity data. Test-retest reliability and validity correlations for the self-report instruments were similar in magnitude to those reported for nonclinical adult samples. Some activity measures were correlated with cognitive status and quality of life. Results indicate that self-reported physical activity can provide valid information for clinical and research applications in mental health settings. Correlations between physical activity measures and psychiatric symptoms suggested interesting patterns for future study.