<p>Acute myeloid leukemia (AML) is a high-risk hematologic malignancy with poor long-term survival and frequent relapse, sustained by leukemic stem cells, antigenic heterogeneity, and an immunosuppressive bone marrow niche. Although chimeric antigen receptor (CAR) T-cell therapy achieves durable responses in B-cell malignancies, its application in AML is restricted by on-target myelotoxicity from antigen overlap with normal progenitors, heterogeneous and dynamic antigen expression, rapid T-cell exhaustion in suppressive microenvironments, limited manufacturing windows with compromised T-cell quality, and uncertainty in optimal infusion timing. To address these barriers, logic-gated and adapter CARs are engineered to broaden antigen recognition while limiting toxicity; nanobody-based CARs provide stable, low-immunogenic binding; gene-edited hematopoietic stem and progenitor cells permit AML clearance without prolonged marrow suppression; and metabolic or epigenetic modulation is employed to sustain T-cell function in hostile niches. Allogeneic CAR-T platforms offer a potential means to overcome manufacturing constraints and improve treatment accessibility. In selected settings, sequential CAR-T therapy and hematopoietic stem cell transplantation consolidate remission and restore hematopoiesis. This review integrates current and emerging AML antigen targets with engineering innovations into a structured translational framework, directly addressing the biological, manufacturing, and application barriers unique to AML, and outlining strategies with the potential to advance CAR-T therapy from experimental studies to durable clinical benefit.</p>