The Beck Anxiety Inventory (BAI) is a widely used anxiety scale in clinical practice as well as in research. In the present study the objective was to assess the psychometric properties of the Icelandic version of the BAI in a patient and a student population. There was a total of 1674 participants, 607 outpatients and 1067 students. All participants completed the BAI and a subgroup of the students completed additional measures of anxiety and depression as well. A subgroup of the students also completed the BAI on a second occasion. Most of the patients completed the Beck Depression Inventory-II (BDI-II). All patients were diagnosed with the Mini-International Neuropsychiatric Interview. Both internal consistency reliability and test-retest reliability were excellent. Convergent and divergent validity were supported. The BAI showed discriminant validity both with regard to discriminating anxiety disorder patients from other patients and panic disorder patients from other anxiety disorder patients. Confirmatory factor analyses revealed excellent support for a one-factor model in the student population, which was superior to the alternative two-and four-factor models. The four-factor model was, however, supported in the patient population whereas the one-and two-factor models were not. It is concluded that the psychometric properties of the Icelandic version of the BAI are satisfactory.