We investigated the effect of psychoanalytically oriented inpatient psychotherapy on three essential forms of anxiety disorders (classified according ICD-10): 1. generalized anxiety disorder (n = 23), agoraphobia (n = 38) and panic disorder (n = 24). Examinations with psychometric test instruments were performed at beginning of inpatient psychotherapy, at dismissal and 6 weeks after dismissal. At the end of inpatient psychotherapy we judged four clinically relevant criteria of therapy success: at the one side the therapist's judgement of symptomatic and structural improvement of the patients, at the other side the patients' judgement of anxiety and depression diminuation. The results demonstrate that summing up these four success criteria 1.) 40.0% of the patients with anxiety disorders improved, 21.2% failed; 2.) 61.4% of the patients with agoraphobia improved, 6.3% failed and 3.) 52.5% of the patients with panic disorder improved, 6.5% failed. An additional analysis of these results showed that the patients with generalized anxiety disorder were significantly (p < 0.01) more disturbed (according clinical assessment) compared to the other two diagnostic groups. About 60% of these patients exhibited increased anxiety symptoms during the course of inpatients psychotherapy, nevertheless relative to the other success criteria these patients achieved satisfactory outcome results. In the paper the consequences of these findings are discussed especially for inpatient psychotherapy of severely disturbed patients with anxiety disorders.