The Diagnostic and Statistical Manual of Mental Disorders (DSM) model of psychopathology has generally been the accepted standard in North America for understanding and diagnosing psychological disorders for over half a century. This classification model, particularly since DSM‐III, has been formulated around the goals of aiding diagnosis, enhancing communication among professionals, fostering psychopathology research and informing treatment. However, all classification systems are inherently dependent on the purpose for the classification. In this paper, an argument is made for a clinically‐relevant diagnostic system of mental disorders to support a primary goal of informing treatment. Several lines of research are examined, including studies on diagnostic reliability, dimensional vs categorical nature of anxiety disorders, co‐morbidity, and psychotherapeutic and pharmacological treatment outcome as they relate to current and proposed diagnostic models of anxiety disorders. Based on the evidence, suggestions are made for revising diagnostic models of anxiety, and key lines of future research are proposed.