Diabetes mellitus (DM) causes dysfunction of endothelial progenitor cells (EPCs), resulting in impaired wound healing. EPC therapy is a potential substitute to the current treatments of chronic wounds. Because EPCs isolated from diabetic patients are dysfunctional and therefore pose an obstacle in their efficacious employment in autologous cell therapy, a strategy to rescue them prior to transplantation would be expected to improve the efficacy of autologous cell therapy multifold. Compromised reactive oxygen species scavenging ability being the main cause of EPC dysfunction (EPCD), reactive oxygen species scavengers are likely to reverse or rescue EPCD. Therefore, in this study, we evaluated the potential of curcumin in reversing DM-induced EPCD. We found that in vitro treatment of bone marrow EPCs from diabetic mice (D-EPC) with curcumin restored their functionality, as judged by colony formation, tubule formation, and migration assays. Most importantly, autologous transplantation of curcumin-treated D-EPCs onto diabetic wounds also resulted in accelerated wound healing. Furthermore, curcumin-treated diabetic mice exhibited improved wound healing, as compared with their vehicle-treated diabetic counterparts, underscoring the efficacy of curcumin in vivo as well. The levels and activity of manganese superoxide dismutase (MnSOD) in D-EPCs treated in vitro with curcumin or those isolated from curcumin-treated diabetic mice were comparable with those in non-diabetic EPCs. Addition of methyl mercury chloride to inhibit MnSOD activity during curcumin treatment abolished the salutary effects of curcumin. Our data demonstrate that curcumin reverses DM-induced EPCD by boosting MnSOD expression and activity and emphasizes its potential for use in autologous cell therapy for diabetic wound management.