Previous studies have shown the potential for intervertebral disc tissue regeneration is very limited. While in vivo and in vitro studies have shown that traction can restore disc height and internal pressure, in many clinical studies it was shown that axial mechanical traction for the treatment of low back pain is ineffective. The aim of this study was to identify how the disc could be distracted, how to define the state of traction, and to further examine the feasibility of regenerating or restoring the degenerative disc by means of traction. A macro- and microlevel structural analysis of degenerative discs of rat tail before and after controlled immobilization-traction. In this study, 49 6-month-old male Sprague-Dawley rats were randomly assigned to one of seven groups. Group A was the sham control group in which caudal vertebrae were instrumented with K-wires only. In Group B (model group), caudal vertebrae were immobilized using a custom-made external device to fix four caudal vertebrae (Co7-Co10) and Co8-Co9 underwent 4 weeks of compression to induce moderate disc degeneration. In Group C, vertebrae Co8-Co9 underwent 4 weeks of compression to induce moderate disc degeneration, followed by removal of the external apparatus. Rats in the other four groups (Groups D-G), Co8-Co9 underwent 4 weeks of compression to induce moderate disc degeneration followed by 2 weeks, 4 weeks, 6 weeks, and 8 weeks of distraction, respectively. Caudal vertebrae were harvested and disc height, T2 signal intensity of the discs, disc morphology, total glycosaminoglycan content of the nucleus pulposus and the structure of the Co8-Co9 end plate were evaluated. After 4 weeks of compression, the intervertebral height and T2 signal intensity of Co8-Co9 vertebrae of rats in Groups B to G were significantly reduced compared with Group A (sham group, all p<.0001). Histological scores of rats in Group B averaged 10.14 and the total glycosaminoglycan (GAG) of nucleus pulposus averaged 238.21μg GAG/ng DNA. The bony end plate structure showed significant changes in comparison with the control Group. After 2 weeks to 8 weeks of traction, the disc space and T2 signal intensity of Co8-Co9 vertebrae in Group E were significantly recovered compared to that of rats in Group B (p<.0001), and the intervertebral height of the Co8-Co9 in Group D, Group F, and Group G when compared with Group B (p<.0001). Meanwhile, the T2 signal intensity of Co8-Co9 in Group D, F, and G when compared with Group B (p<.001). Histological scores dropped from an average of 10.14 in Group B to 5.57 in Group E, and 5.86 in Group F (all p<.0001). Furthermore, the total GAG content of the nucleus pulposus increased from an average of 238.21μg GAG/ng DNA in Group B to 601.02μg GAG/ng DNA in Group E (p<.0001). The number of pores of end plates in rats in Groups D and E both were significantly increased when compared to that of rats in Group B (Groups D vs Groups B, p<.05; Groups E vs Groups B, p<.0001). A mechanical degenerative model was successfully established by using a custom-made device. We demonstrated that disc degeneration is a cascade of biochemical, mechanical, and structural changes mediated by cells in an abnormal mechanical environment. Not all levels of disc degeneration can be regenerated or repaired. Regeneration or recovery of disc degeneration requires specific conditions. Based on the immobilization-traction mode, the cascade cycle of disc degeneration is interrupted. Traction of 2 to 6 weeks is a sensitive period for regeneration of the degenerative disc. Moreover, the duration and extent of the traction loading must be moderately controllable, and beyond the limits that can lead to significant degeneration. These data may help improve our understanding of the pathogenesis of clinical disc degeneration and how to optimize the use of traction devices for possible regeneration.