Lentiviral vector forkhead box P3 (Foxp3)/pXZ9 containing Foxp3-IRES-GFP fragment and its mock control pXZ9 was constructed. Lentiviruses were produced via transient 3-plasmid transfection. BALB/c CD4+CD25−T cells were infected with lentiviruses and further stimulated using anti-CD3x3b5; and anti-CD28 antibody (engineered irrelevant-Tregs [irTregs]) or C57BL/6 splenocytes (engineered sTregs). The expression of Tregs marks, production of cytokines, cell proliferation rate, and suppression function of Foxp3/pXZ9 infected cells were similar to natural Tregs. Irradiation BABL/c recipient were injected with C57BL/6 donor T cell–depleted bone marrow (TCD-BM) cells (1 × 107) and C57BL/6 splenocytes (1 × 107) together with engineered sTregs, irTregs, or natural Tregs (5 × 106). Irradiated BABL/c mice received TCD-BM cells only, TCD-BM cells plus splenocytes, or splenocytes and pXZ9-transduced cells (control). Recipient survival, short-term GvHD scores, and the Th1 subpopulation were monitored. Recipients of a combination of TCD-BM cells and splenocytes developed lethal GVHD. When engineered sTregs were added, 80 % of recipients survived at least 60 days after transplantation; this survival rate was much higher than in any other group. The GvHD scores between the 3 Tregs groups did not demonstrate significance. Compared with other sources of Tregs in vivo, engineered sTregs strongly suppressed Th1 cell expansion. Therefore, a an in vitro strategy was developed to generate engineered sTregs. These cells demonstrated similar phenotypes and stable suppressive capacity as natural Tregs. Like natural Tregs, co-injection of engineered Tregs protected recipients from lethal GvHD in a murine model of GvHD. The engineered sTregs were superior to irTregs in minimizing murine GvHD.