AIM: Some researches show that stem cell transplantation for damaged spinal cord can improve the function of damaged spinal cord. But the studies about bone marrow mononuclear cell transplantation for injuried spinal cord are seldom. We transplanted fresh bone marrow mononuclear cells isolated from rats into rat models of injured spinal cord to explore the effect of bone marrow mononuclear cells for injured spinal cord functions, nerve regeneration, neovascuarization and long-term effect. METHODS: Experiments were performed in the Experiment Center of Developmental Biology of Shanghai Second Medical University from October 2005 to April 2006. The laboratory is Specific-pathogen free grade. ①Female clean SD rats aged 8 weeks weighting 200-220 g were offered by Animal Experimental Centre of Chinese Academy of Sciences. Animal intervention met the animal ethical standard. ②Rat bone marrow mononuclear cells were isolated from the tibia and the femur by Ficoll-Paque density gradient centrifugation. Rat models of spinal injury were established. The 22 successfully established rat models were divided into 2 groups. Rat models in a model plus cell group (n =11) received the complete T9-10 transection of spinal cord, and then bone marrow mononuclear cells were transplanted into the vertebral canal. Rat models in a model plus DMEM group (n =11) received the complete T9-10 transection of spinal cord, and then DMEM was injected into adjacent region. Rat models in a sham operation group (n =9) received T9-10 spinous process and lamina of vertebra incision and then the incision was sutured. ③Hybridization in situ and immunohistochemistry technique were used to determine the survival of implanted cells in host spinal cord. BBB scale system was applied to assess the functional recovery of spinal cord nerves. RESULTS: ①There was no significant difference in postoperative score at each time point in the sham operation group. The score was 0 point in the model plus DMEM group. The function of spinal cord did not recover. The function of spinal cord became better at weeks 2, 4, 6 and 8 in the model plus cell group. There were significant differences as compared with the model plus DMEM group and the sham operation group (P