Short screening questionnaires for mental health are useful tools for research and clinical practice, e.g. they could play a major role in detecting patients with psychiatric disorders in primary care. The present study tests the validity of the five-item Mental Health Inventory (MHI-5) screening test using DSM-IV Axis I diagnoses as a gold standard and analyzes its performance in different diagnostic groups. A random sample was drawn from the resident registration office files in northern Germany. Personal interviews with a response rate of 70% were conducted. Of the sample, 4036 respondents filled in the MHI-5. DSM-IV diagnoses were assessed using the Munich Composite International Diagnostic Interview (M-CIDI). The area under the receiver operating characteristics curve (AUC) of 0.72 in identifying any DSM-IV Axis I disorder (except substance use) is not satisfying. The MHI-5 revealed best performance for mood (AUC: 0.88) followed by anxiety disorders (AUC: 0.71). Sensitivity and specificity were poor for somatoform and substance use disorders, especially in cases without comorbid mood or anxiety disorder. The power to detect mood and anxiety disorders of the MHI-5 was better for the 4-week compared with the 12-month diagnoses. The MHI-5 can be recommended to screen for mood disorders. Data have to be confirmed for primary care settings.