We systematically review the literature on approach-avoidance (AA) tendencies in mental disorders, including 97 empirical studies. Most evidence for the role of biased AA tendencies was found in addictive disorders: The presence of an approach bias (ApB) for substance related stimuli in subclinical populations can be a risk factor for increased future substance use, and AA modification training given as an add-on to standard treatment has the potential to reduce intake and relapse rates reliably. In depression, reduced approach of positive stimuli and reduced avoidance of negative stimuli have been found, and modification procedures seem to have clinical potential. In anxiety disorders, an avoidance bias (AvB) for threat-related stimuli has been found frequently, but modification studies did not yield any clinical effects. In eating disorder a lack of food preferences in anorexia nervosa may be present, but relations between AA measures and clinical (outcome) measures were not established. In other disorders, the evidence was limited due to a low number of published studies. Several methodological problems are discussed: It is often difficult to compare studies to each other, control groups and control stimuli are frequently missing, and many studies suffer from insufficient statistical power due to small samples. We finally give suggestions for future research on biased AA tendencies in psychopathology.