Background Although there is evidence for the effectiveness of interventions for psychosis among ultra-high-risk (UHR) groups, health economic evaluations are lacking. This study aimed to determine the cost effectiveness and cost-utility of cognitive-behavioural therapy (CBT) to prevent first-episode psychosis. Method The Dutch Early Detection and Intervention Evaluation study was a randomized controlled trial of 196 UHR patients with an 18-month follow-up. All participants were treated with routine care (RC) for non-psychotic disorders. The experimental group (n = 95) received add-on CBT to prevent first-episode psychosis. We report the intervention, medical and travel costs, as well as costs arising from loss of productivity. Treatment response was defined as psychosis-free survival and quality-adjusted life years (QALYs) gained. Results In the cost-effectiveness analysis, the proportion of averted psychoses was significantly higher in the CBT condition (89.5% v. 76.2%). CBT showed a 63.7% probability of being more cost effective, because it was less costly than RC by US844 (£551) per prevented psychosis. In the cost-utility analysis, QALY health gains were slightly higher for CBT than for RC (0.60 v. 0.57) and the CBT intervention had a 52.3% probability of being the superior treatment because, for equal or better QALY gains, the costs of CBT were lower than those of RC. Conclusions Add-on preventive CBT for UHR resulted in a significant reduction in the incidence of first psychosis. QALY gains show little difference between the two conditions. The CBT intervention proved to be cost saving.