Transplantation-related mortality (TRM) is high after HLA-mismatched umbilical cord blood (UCB) transplantation (UCBT). In utero, exposure to noninherited maternal antigen (NIMA) is recognized by the fetus, which induces T regulator cells to that haplotype. It is plausible that UCBTs in which recipients are matched to donor NIMAs may alleviate some of the excess mortality associated with this treatment. To explore this concept, we used marginal matched-pair Cox regression analysis to compare outcomes in 48 NIMA-matched UCBTs (ie, the NIMA of the donor UCB unit matched to the patient) and in 116 non-NIMA-matched UCBTs. All patients had a hematologic malignancy and received a single UCB unit. Cases and controls were matched on age, disease, disease status, transplantation-conditioning regimen, HLA match, and infused cell dose. TRM was lower after NIMA-matched UCBTs compared with NIMA-mismatched UCBTs (relative risk, 0.48; P = .05; 18% versus 32% at 5 years posttransplantation). Consequently, overall survival was higher after NIMA-matched UCBT. The 5-year probability of overall survival was 55% after NIMA-matched UCBTs versus 38% after NIMA-mismatched UCBTs (P = .04). When faced with the choice of multiple HLA-mismatched UCB units containing adequate cell doses, selecting an NIMA-matched UCB unit may improve survival after mismatched UCBT.