Endothelium dysfunction is a central problem for early rejection due to the host alloimmune response and the late status of arteriosclerosis in heart transplantation. However, reliable pieces of evidence are still limited concerning the source of the regenerated endothelium within the transplanted heart. We analyzed single-cell RNA sequencing data and constructed an inducible lineage tracing mouse, combined heart transplantation with bone marrow transplantation and a parabiosis model, cellular components, and endothelial cell populations in cardiac graft lesions. Our single-cell RNA sequencing analysis of a transplanted heart allowed for the establishment of an endothelial cell atlas with a heterogeneous population, including arterial, venous, capillary, and lymphatic endothelial cells. Along with genetic cell lineage tracing, we demonstrated that the donor cells were mostly replaced by recipient cells in the cardiac allograft, up to 83.29% 2weeks after transplantation. Furthermore, recipient nonbone marrow CD34+ endothelial progenitors contributed significantly to extracellular matrix organization and immune regulation, with higher apoptotic ability in the transplanted hearts. Mechanistically, peripheral blood-derived human endothelial progenitor cells differentiate into endocardial cells via Vascular endothelial growth factor receptor-mediated pathways. Host circulating CD34+ endothelial progenitors could repair the damaged donor endothelium presumably through CCL3-CCR5 chemotaxis. Partial depletion of host CD34+ cells resulted in delayed endothelial regeneration. We created an annotated fate map of endothelial cells in cardiac allografts, indicating how recipient CD34+ cells could replace the donor endothelium via chemokine CCL3-CCR5 interactions. The mechanisms we discovered could have a potential therapeutic effect on the long-term outcomes of heart transplantation.