Abstract Aim Capecitabine and bevacizumab (CAP-B) maintenance therapy has shown to be more effective compared with observation in metastatic colorectal cancer patients achieving stable disease or better after six cycles of first-line capecitabine, oxaliplatin, bevacizumab treatment in terms of progression-free survival. We evaluated the cost-effectiveness of CAP-B maintenance treatment. Methods Decision analysis with Markov modelling to evaluate the cost-effectiveness of CAP-B maintenance compared with observation was performed based on CAIRO3 study results (n = 558). An additional analysis was performed in patients with complete or partial response. The primary outcomes were the incremental cost-effectiveness ratio (ICER) defined as the additional cost per life year (LY) and quality-adjusted life years (QALY) gained, calculated from EQ-5D questionnaires and literature and LYs gained. Univariable sensitivity analysis was performed to assess the influence of input parameters on the ICER, and a probabilistic sensitivity analysis represents uncertainty in model parameters. Results CAP-B maintenance compared with observation resulted in 0.21 QALYs (0.18LYs) gained at a mean cost increase of €36,845, yielding an ICER of €175,452 per QALY (€204,694 per LY). Varying the difference in health-related quality of life between CAP-B maintenance and observation influenced the ICER most. For patients achieving complete or partial response on capecitabine, oxaliplatin, bevacizumab induction treatment, an ICER of €149,300 per QALY was calculated. Conclusion CAP-B maintenance results in improved health outcomes measured in QALYs and LYs compared with observation, but also in a relevant increase in costs. Despite the fact that there is no consensus on cost-effectiveness thresholds in cancer treatment, CAP-B maintenance may not be considered cost-effective.