Objectives: Pathophysiological mechanisms underlying spinal cord injury (SCI) partially involve edema and formation of a hematoma. Myelotomy seems to be a promising intervention. However, the appropriate timing of myelotomy is still unknown in SCI. Here we aimed to determine the timing of microsurgical myelotomy in an animal model of SCI. Methods: The SCI model was contusion-induced with a new york university impactor. Sixty-five adult female rats were randomly divided into the following groups: laminectomy alone (the 'sham group', SG), laminectomy plus contusion (the 'contusion group', CG) or laminectomy plus contusion followed by myelotomy at 8, 24 or 48 h (8 h-MTG [myelotomy-treated group], 24 h-MTG or 48 h-MTG). Functional recovery was evaluated via the open field test and the inclined plane test every week after SCI. The percentage of spared white matter area (SWMA) and ultrastructure characteristics of the injured dorsolateral spinal cord were determined on the 42nd day after SCI. Results: Compared with the CG, myelotomy at 8 h-MTG or 24 h-MTG greatly improved the BASSO-BEATTIE-BRESNAHAN scores (Po0.008), whereas the 48 h-MTG showed less efficacy (P ¼ 0.023). All myelotomy groups showed higher mean angle values in an inclined plane test (Po0.005) and had greater percentages of SWMA than the CG. Rats in the 24 h-MTG showed a higher intraaxonal fraction and myelin fraction than those in 48 h-MTG (Po0.005). Conclusion: Myelotomy up to 48 h after SCI improves recovery in rats. The potential time window of myelotomy may be between 8 and 24 h after SCI.