We examined the utility of the Affective States Questionnaire (ASQ) in predicting acute risk for suicidal behavior. Subjects at a VHA Medical Center were interviewed using the ASQ and again 3 months later when their suicidal behaviors over that period were examined. The ASQ had a sensitivity of 60% for predicting suicidal behavior over the follow-up period, and specificity of 74%. The false positive rate was relatively low for a sample not highly selected for suicide risk and utilizing a short period of 3 months for suicidal behavior. Subgroups combining the ASQ with disability level or a diagnosis of substance abuse greatly reduced the percentage of false positives. The ASQ is able to improve significantly our ability to predict acute risk of suicidal behavior in clinical psychiatric populations.