This prospective randomized study compared 3 fusion methods: posterolateral fusion (PLF), posterior lumbar interbody fusion (PLIF), and PLIF combined with PLF (PLF+PLIF). To compare the outcomes of the 3 fusion methods and find a useful fusion method. Many studies have shown clinical results, advantages, and postoperative complications of each fusion method, but few have compared the 3 fusion methods prospectively. A total of 167 patients who underwent 1 or 2-level fusion surgery because of degenerative lumbar disease from January 1996 to September 2000 were studied. Minimum follow-up was 3 years. The patients were randomized into 1 of 3 treatment groups: group 1 (PLF; n = 62); group 2 (PLIF; n = 57); and group 3 (PLF+PLIF; n = 48). A visual analog scale, the Oswestry Disability Questionnaire, and Kirkaldy-Willis criteria were used to measure low back pain, leg pain, and disability. For radiologic evaluation, disc height, lumbar lordosis, segmental angle, and bone union were examined. Postoperative complications were also analyzed. At the last follow-up, good or excellent results were obtained in 50 cases of PLF (80.7%), 50 cases of PLIF (87.8%), and 41 cases of PLF+PLIF (85.5%). No statistical differences were found among the 3 groups (P = 0.704). All methods indicated significant improvement in the disc height (P 0.05). Complications included deep infection in 3 cases, transient nerve palsy in 4, permanent nerve palsy in 1, and donor site pain in 6. No significant differences in clinical results and union rates were found among the 3 fusion methods. PLIF had better sagittal balance than PLF. PLIF without PLF had advantages of the elimination of donor site pain, shorter operating time, and less blood loss.