Objective To establish animal models of spinal cord injury of different degrees of severity, namely,MASCIS and IH, and to study the relationship among the degree of damage, modeling mode, behavior, and anatomical results. Methods Eighty female SD rats were randomly and equally divided into four groups: NYU 12. 5 mm, NYU 25mm, IH 150 kdyn force, and IH 200 kdyn force models. Their rates of successful model establishment, BBB score, grid walking evaluation, Hargreaves test, and total cross-sectional area of the spinal cord at the injury epicenter were evaluated.Results The success rates in the two modeling methods were similar. The BBB scores of all animals were reduced to 0points at 0 to 3 days after spinal cord injury. As time passed, the BBB scores gradually recovered and reached a plateau at 4to 6 weeks. The NYU 25 mm group scored the lowest, the IH 200 group had an intermediate score, and the IH150 and NYU 12. 5 mm groups scored similarly and highest. The grid walking score increased with increasing injury severity. The Hargreaves test results were more affected by motor function. Dyskinesia was inversely proportional to the residual amount of tissue at the center of the injury. Conclusions Both MASCIS and IH models are appropriate for research on the pathophysiological mechanisms of spinal cord injury, but appropriate modeling parameters should be selected according to the purpose of the experiment.