This study aims to evaluate the long-term cost-effectiveness of add-on cognitive behavior therapy (CBT) for the prevention of psychosis for individuals at ultrahigh risk (UHR) of psychosis. The Dutch Early Detection and Intervention randomized controlled trial was used, comparing routine care (RC; n = 101) with routine care plus CBT for UHR (here called CBTuhr; n = 95). A cost-effectiveness analysis was conducted with treatment response (defined as proportion of averted transitions to psychosis) as an outcome and a cost-utility analysis with quality-adjusted life years (QALYs) gained as a secondary outcome. The proportion of averted transitions to psychosis was significantly higher in the CBTuhr condition (with a risk difference of 0.122; b = 1.324, SEb = 0.017, z = 7.99, P < 0.001). CBTuhr showed an 83% probability of being more effective and less costly than RC by -US$ 5777 (savings) per participant. In addition, over the 4-year follow-up period, cumulative QALY health gains were marginally (but not significantly) higher in CBTuhr than for RC (2.63 vs. 2.46) and the CBTuhr intervention had a 75% probability of being the superior treatment (more QALY gains at lower costs) and a 92% probability of being cost-effective compared with RC at the Dutch threshold value (US$ 24 560; €20 000 per QALY). Add-on preventive CBTuhr had a high likelihood (83%) of resulting in more averted transitions to psychosis and lower costs as compared with RC. In addition, the intervention had a high likelihood (75%) of resulting in more QALY gains and lower costs as compared to RC.