We present a prospective, randomized study of acute burst fracture of the thoracolumbar spine. Forty-one patients were treated either by anterior decompression and stabilization with Kaneda device or by posterior distraction instrumentation using the A-O fixateur interne. The mean follow up was 21 months. The result were as follows; 1. The mean preoperative kyphotic angle was 19.2° in those patients treated by anterior surgery and 21.4° in those patients treated by posterior surgery. At last follow-up the mean correction in kyphotic angle was 13.2° in the anterior group and 9.2° in the posterior group. There is no statistically significant difference between those two groups. 2. The mean preoperative midsagittal diameter of the canal compromise 47.4% in anterior group and 49.3% in posterior group. Postoperatively, this was reduced to 2.5% and 7.9%. There is a statistically significant difference between these two groups(P<0.05). 3. The mean preoperative canal enchroachment 52.)% in anterior group and 47.6% in posterior group. Postoperatively, this was reduced to 3.2% and 6.0%. There is a statistically significant difference between these two group(P<0.05). 4. Neurologic recovery was 81% in posterior surgery and 85% in anterior surgery. The improvement in Frankel grade was 1 grade in average, and showed no difference between two groups. 5. There was two cases of pedicle screw breadage in posterior group and one case of pyothorax in anterior group but no early or late vascular or neurologic complication.