Background It remains almost a helpless situation for the recurrent implantation failure and pregnancy loss caused by endometrial injury at present. The purpose of this study was to develop a rabbit model of endometrial mechanical injury that could provide a research platform for this difficult clinical predicament.Methods Three experiments were conducted. Experiment 1 : Curettages in both uterus horns and copper wire inserting after curettage (double-injury) in one horn. The histological changes were monitored at 0, 24, 48, 72 hours, as well as in 1 and 2 weeks after operation. Experiment 2: Direct copper wire inserting in one horn and double-injury in other horn. The wires in both horns were removed after 2 weeks. The histological changes were recorded at 0, 1 and 2 weeks after wire removal. Experiment 3: Double-injury procedure in one horn was performed and wire was removed after 2 weeks;another horn was remained normal to serve as control. Histological changes were recorded, tissue areas were measured,and proliferation indices (PIs, %) were calculated at 1,2, 4 and 8 weeks after wire removal, respectively.Results The experiments revealed that the injured endometrium by simple curettage or copper wire could be fully repaired. While the endometrial regeneration was severely impaired by double-injury, both areas of endometrium and uterine cavity decreased (P ＜0.05); both PIs of glandular epithelial and stromal cells increased and reached maximum at 4 weeks (P ＜0.05), but returned by 8 weeks.Conclusion This study demonstrated that a rabbit model of endometrial injury could be effectively established through a double-injury procedure of curettaqe and copper wire with comparable clinical index.