This study was performed to investigate incidence, causes and factors influencing mortality after haploidentical hematopoietic stem cell transplantation (HSCT) and to compare differences between haploidentical HSCT and HLA-identical sibling HSCT. From January 2000 to June 2011, 1411 patients with acute leukemia or myelodysplastic syndrome were included in this study. Of these patients, 571 received HLA-identical sibling HSCT and 840 received haploidentical HSCT. The cumulative incidence of overall mortality and transplant-related mortality (TRM) after haploidentical HSCT was higher than those after HLA-identical sibling HSCT (38.7% vs 33.3%, P = 0.012 and 27.5% vs 19.9%, P = 0.002), but the incidence of relapse-related mortality (RRM) did not differ between the two groups (15.6% vs 16.7%, P = 0.943). A multivariate analysis suggested that high-risk disease status and haploidentical HSCT correlated with a higher incidence of overall mortality (P o0.0001, hazard ratio = 1.911 and P = 0.019, hazard ratio = 1.249); in addition, in haploidentical HSCT, only high-risk disease status correlated with a higher incidence of overall mortality (P o0.0001, hazard ratio = 1.845). Our study suggested that haploidentical HSCT provided a higher incidence of overall mortality and TRM but the same incidence of RRM compared with HLA-identical sibling HSCT. Therefore, HLA-identical sibling HSCT remains the first choice, but haploidentical HSCT is available for patients without an HLA-identical sibling donor.