Introduction Alcohol use disorder (AUD) is a brain disorder linked to over 200 health conditions. Cognitive behavioural therapy (CBT) is considered the best practice in the treatment of AUD, but more than 60% of patients relapse within the first year after treatment. Psychotherapy combined with virtual reality (VR) has received increasing interest in the treatment of AUD. However, existing studies have primarily investigated the use of VR for cue reactivity. We therefore aimed to investigate the effect of VR-assisted CBT (VR-CBT). Methods and analysis This study is an assessor-blinded, randomised clinical trial being conducted at three outpatient clinics in Denmark. We will randomise 102 patients to 14 individual sessions of either manualised VR-CBT or CBT. The VR-CBT group will receive exposure to immersive high-risk VR situations from a pub, bar/party, restaurant, supermarket and at-home (30 videos) to activate high-risk-related beliefs and cravings for subsequent modification using CBT techniques. The treatment period is 6 months, and follow-up visits will be performed 3, 6, 9 and 12 months after inclusion. The primary outcome measure is the change in total alcohol consumption from baseline to 6 months after inclusion, measured with the Timeline Followback Method. Key secondary outcome measures include changes in the number of heavy drinking days, alcohol cravings, cognition, and symptoms of depression and anxiety. Ethics and dissemination Approval has been obtained by the research ethics committee in the Capital Region of Denmark (H-20082136) and the Danish Data Protection Agency (P-2021-217). All patients will receive both oral and written information about the trial and written informed consent will be obtained from each patient before inclusion. The study results will be disseminated in peer-reviewed publications and conference presentations. Trial registration number ClinicalTrial.gov, NCT05042180.