Objective：To compare curative effect between anterior decompression and posterior decompression in the treatment of tho－racolumbar burst fracture with neurologic deficit. Methods：From March 2007 to March 2010，62 cases of thoracolumbar burst fracture with neurologic deficit were treated. There were 43 males and 19 females with an average age of 42.3 years（22-68 years）. All cases were randomly divided into 2 groups：group A（anterior spinal canal decompression） and group B （posterior spinal canal decompres－sion）. According to American Spinal Injury Association（ASIA） standard，group A had 19 cases of C and 12 cases of D；group B had 16 cases of C and 15 cases of D. Results：Totally 62 cases were followed up for 2.0-3.5 years（28 months in average）. All wounds achieved primary healing postoperatively and all cases achieved satisfactory results in vertebral canal decompression and reduction. No loose－ness，protrution and breakage of internal fixator was observed. Blood loss was （675.3±124.1） ml in group A and （381.3±77.8） ml in group B，with statistically significant differences（t=11.175，P=0.000）. Operative time was （193.68±22.01） min in group A and （120.51±19.83） min in group B，with statistically significant differences（t=13.748，P=0.000）. Preoperative vertebral canal volume in group A was （43.05±8.53）% and （45.29±7.60）% in group B，without statistically significant differences（t=-1.092，P=0.279）. Post－operative vertebral canal volume in group A was （91.50±2.91）% and （89.81±2.60） % in group B，with statistically significant dif－ferences（t=2.411，P=0.010）. Vertebral canal volume in the last follow-up in group A was （91.89±2.16）% and （90.90±1.86）% in group B，with statistically significant differences（t=1.934，P=0.029）. Preoperative，postoperative and final vertebral heights in group A were （43.97±8.26）%，（90.45±2.26）%，（89.67±2.09）% and （45.48±6.52）%，（89.63±2.16）%，（88.95±1.85）% in group B，with－out statistically significant differences（t=-0.799，t=1.460，t=1.436，P＞0.05）. At 2 years postoperatively，27 cases in group A partly re－covered（curative rate of 87.10%） and 24 cases in group B partly recovered（curative rate of 77.45%），showing no statistically signifi－cant differences（χ2=0.995，P=0.800）. There were statistically significant differences in the postoperative pain，medical cost and time in hospital between groups A and B（t=2.181，t=11.808，t=5.848，P＜0.05）. Conclusions：Posterior decompression can achieve the same effects as anterior decompression in the treatment of thoracolumbar burst fractures with neurologic deficit as long as indications are strictly controlled. Posterior decompression is easier and has more merits than anterior decompression：shorter operative time，less blood loss，less damage，lower medical cost. Posterior decom－pression is more suitable for hospitals at primary level.