Objective To investigate the technique of reduction by posterior approach for severe spondylolisthesis, and to discuss the method to prevent nerve stretch injury. Methods Between July 2007 and April 2011, 17 patients with severe spondylolisthesis underwent reduction, fixation, and fusion by posterior approach. There were 2 males and 15 females with a median age of 15 years (range, 8-67 years) and a median disease duration of 18 months (range, 5 months-16 years and 4 months). The level of spondylolisthesis was at L4 in 1 case and L5 in 16 cases; the spondylolisthesis was at degree III in 12 cases and degree IV in 5 cases according to Meyerding classification. There were 16 cases of developmental spondylolisthesis (high- dysplastic and low-dysplasia spondylolisthesis in 9 and 7 cases, respectively) and 1 case of traumatic spondylolisthesis; 16 cases of developmental spondylolisthesis at L5 level included 6 cases of type 4, 9 case of type 5, and 1 case of type 6 according to Spinal Deformity Study Group (SDSG) classification. All cases underwent posterior spinal decompression, Schanz screw fixation for the slipped vertebrae, the intervertebral and posterolateral fusion and reduction of the slipped vertebrae, and correction of the lumbosacral kyphosis. The reductive degree of slipped vertebrae was modulated according to the strain of exiting spinal root. The slip degree should be reduced within Meyerding degree II. The anteroposterior and lateral radiographs of whole spine were taken in a standardized standing position to observe the correction of displacement severity and lumbosacral angle. The nerve function and pain score of lower extremity were evaluated by neurological Frankel grade and visual analogue scale (VAS). Bony fusion was assessed by followed-up CT three-dimentional reconstruction. Results Exiting nerve root paralysis occurred in 1 case after operation, and released at 4 weeks after operation; no aggravation of nerve damage was observed in the other patients. The incisions primarily healed. All the patients were followed up 12-48 months (mean, 25 months). The slip percentage, the lumbosacral angle, and VAS score of lower extremity were improved from 72% ± 10%, (18.2 ± 3.5)°, and 7.0 ± 1.5 at preoperation to 12% ± 6%, ( — 7.3 ± 2.9)°, and 1.5 ± 1.3 at 12 months after operation respectively, all showing significant differences (P lt; 0.05). Osteosynthesis was seen at the bone grafting area by CT three-dimentional reconstruction at 12 months after operation. No breakage of screw and rod or reduction loss occurred. Conclusion It can obtain satisfactory clinical result to use spinal canal decompression by posterior approach, the Schanz screw fixation of the slipped vertebrae, the intervertebral and posterolateral fusion for severe spondylolisthesis. The risk of nerve stretch injury can be prevented by choosing the lowest height of intervertebral cage, modulating the reductive degree of slipped vertebrae according to the strain of exiting spinal root, and correcting lumbosacral kyphosis.