This study provides psychometric data for the Addiction Severity Index (ASI) when administered to persons with severe and persistent mental disorders. Participants were 97 outpatients (26 women) at a public psychiatric facility. The internal consistency of the composite scores was lower in this psychiatric sample than in previous nonpsychiatric samples. Interrater reliability was acceptable for most composite scores but low for many severity ratings. Several scores showed low temporal stability. Validity evidence was weak for the employment and family-social subscales, acceptable for drug and alcohol subscales, and mixed for psychiatric, medical, and legal subscales. Due to mixed reliability and validity evidence, caution should be exercised when using the ASI with patients having severe mental illness.