Retrospective studies have shown similar survival among patients with acute myeloid leukemia (AML) and myelodysplasia (MDS) after nonmyeloablative compared with myeloablative conditioning. Refined risk stratification is required to design prospective trials. We stratified outcomes among patients with AML (n = 391) or MDS (n = 186) who received either nonmyeloablative (n = 125) or myeloablative (n = 452) allogeneic hematopoietic cell transplantation (HCT) based on comorbidities, as assessed by a HCT-specific comorbidity index (HCT-CI), as well as disease status. Patients receiving nonmyeloablative conditioning were older, more frequently pretreated, more often received unrelated grafts, and more often had HCT-CI scores of 3 compared with patients who received myeloablative conditioning. Patients with HCT-CI scores of 0 to 2 and either low or high disease risks had probabilities of overall survival at 2 years of 70% and 57% after nonmyeloablative conditioning compared with 78% and 50% after myeloablative conditioning, respectively. Patients with HCT-CI scores of 3 and either low or high disease risks had probabilities of overall survival of 41% and 29% with nonmyeloablative conditioning compared with 45% and 24% with myeloablative regimens, respectively. After adjusting for pretransplantation differences, stratified outcomes were not significantly different among patients receiving nonmyeloablative compared with myeloablative conditioning, with the exception of lessened nonrelapse mortality (hazard ratio, 0.50; P = .05) in the highest risk group. Patients with low comorbidity scores could be candidates for prospective randomized trials comparing nonmyeloablative and myeloablative conditioning regardless of disease status. Additional data are required for patients with low-risk diseases and high comorbidity scores. Novel antitumor agents combined with nonmyeloablative HCT should be explored among patients with high comorbidity scores and advanced disease.