Objective: Poor cognitive flexibility has been highlighted in patients with anorexia nervosa (AN), contributing to the development and maintenance of symptoms. The aim of the present study is to investigate how enhanced cognitive flexibility is involved in treatment outcomes in patients with AN. Method: One hundred thirty female out-patients treated for AN have been assessed at baseline and after 4 months of treatment. Path analyses were used to investigate the mediating role of cognitive flexibility, measured through the Brixton test, on a wide range of outcomes: body mass index, eating disorder symptoms, daily life functioning, anxiety, depression, emotions, self-rated silhouette. Results: Cognitive flexibility was improved during treatment, and enhanced cognitive flexibility explains a significant part of level of the improvement in daily life functioning (26%), reduction of eating disorder symptoms (18%) and reduction of depressive symptoms (17%). Others outcomes were also improved, but these improvements were not mediated by cognitive flexibility. Conclusions: Results suggest that enhancing cognitive flexibility could help reduce rigid cognitive and behavioural patterns involved in AN, thus improving everyday functioning and clinical severity. Further studies combining different types of cognitive flexibility evaluation as well as neuroimaging may be necessary to better establish which of its aspects are involved in patients' improvement.