Patients with anorexia nervosa (AN) display elevated anxiety and attention biases (ABs) in threat processing. Attention bias modification treatment (ABMT) is considered promising for anxiety disorders, but its potential for AN is limited. In this study, 154 young women hospitalised because of AN were assigned to ED-related and anxiety-related threat stimuli, or to a non-ABMT intervention control condition in a randomized control trial. Hundred-andten patients completed the study. ABMT was an add-on to the regular inpatient treatment. Research participants completed two pretreatment training sessions and eight biweekly sessions of ABMT. AB, ED-related symptoms, depression, anxiety and stress were assessed before and after ABMT in the research groups, and, similarly, 5 weeks apart, in the controls. We found that despite the different patterns of change in AB between the three groups following ABMT, the reduction in AB, or the between-group differences in AB-reduction, were not significant. While the severity of ED-symptoms, depression, anxiety and stress was reduced following ABMT, or control condition, in all groups, there were no between-group differences in these changes. Changes in AB were not correlated with baseline and pre-posttreatment changes in ED-related and comorbid symptomatology. Methodological and inpatient treatment-related considerations may explain our negative ABMT-related results.