Abstract Higher rate of nonrelapse mortality (NRM) remains yet to be resolved in umbilical cord blood transplantation (UCBT). Considering that UCBT has some unique features compared with allogeneic hematopoietic cell transplantation from other graft sources, a UCBT-specific NRM risk assessment system is required. Thus, in this study, we sought to develop a UCBT-specific NRM Risk Assessment (CoBRA) score. Using a nationwide registry database, we retrospectively analyzed 4437 recipients who had received their first single-unit UCBT. Using the backward elimination method, we constructed the CoBRA score in a training cohort (n = 2687), which consisted of recipients age ≥55 years (score 2), hematopoietic cell transplantation–specific comorbidity index ≥3 (score 2), male recipient, graft-versus-host disease prophylaxis other than tacrolimus in combination with methotrexate, performance status (PS) 2 to 4, HLA allele mismatch ≥ 2, refined Disease Risk Index high risk, myeloablative conditioning, and CD34+ cell doses < 0.82 × 105/kg (score 1 in each). The recipients were categorized into 3 groups: low (0-4 points), intermediate (5-7 points), and high (8-11 points) groups according to the CoBRA score. In the validation cohort (n = 1750), the cumulative incidence of NRM at 2 years was 14.9%, 25.5%, and 47.1% (P < .001), and 2-year overall survival (OS) was 74.2%, 52.7%, and 26.3% (P < .001) in the low, intermediate, and high groups, respectively. In summary, the CoBRA score could predict the NRM risk as well as OS after UCBT. Further external validation will be needed to confirm the significance of the CoBRA score.