【Abstract】 Objective To investigate the manufacture and biocompatibil ity of a bioabsorbable poly-D,L-lactic acid(PDLLA) plate containing rhBMP-2. Methods rhBMP-2 was composited with PDLLA (0.05 mg/plate) through vacuum to repare PDLLA plate containing rhBMP-2. Thirty-two New Zealand rabbits (male or female) weighted (3.0 ± 0.5) kg were used in he study. A 2.5 mm middle ulna osteotomy was made bilaterally. The bones as well as periosteum were removed. The right side of all the animals was experimental side (n=32), was fixed internally by PDLLA plate containing rhBMP-2.The left side of all the animals was control side (n=32), was fixed by common PDLLA plate. After a follow-up of 2, 4, 8 and 12 weeks, the ulnas were examined visually, radiographically, histologically, and by computer graph analysis to compare the biocompatibil ity. Re sults Po rosity of PDLLA plate containing rhBMP-2 was 90%, aperture was 150 μm, tensile strength was higher than 50 MPa, three point flexural strength was higher than 90 MPa and intrinsic viscosity was 1.6 dL/g (chloroform solvent). All animals had a good heal ing 1 or 2 weeks after operation. All the animal’s diet and movement were normal. All the fractures were stable. The plates in the experimental group degraded faster than those in the control group. Relative values of callus density evaluated by X-ray at 2, 4, 8 and 12 weeks after operation in the experimental group were 39.22 ± 2.48, 48.79 ± 1.26, 63.78 ± 1.78 and 78.60 ± 1.25 respectively. Those in the control defects were 33.83 ± 1.13, 41.28 ± 1.25, 55.23 ± 0.68 and 66.54 ± 1.33. At each time point, the experimental side produced better and more trabeculae than the control side did (P ＜ 0.01). Histological examination showed that the newbone formation in the experimental side at 2, 4, 8 and 12 weeks after operation was 0.106% ± 0.015%, 0.292% ± 0.019%, 0.457% ± 0.048% and 0.601% ± 0.037%, while those in the control side was 0, 0.193% ± 0.019%, 0.339% ± 0.029% and 0.574% ± 0.047%respectively. At each time point, the experimental side produced better new bone formation than the control side did (P lt; 0.05). The experimental side showed better compatibil ity between plates and surrounding tissue, faster bone formation, more bone regeneration mass and better medullary canal structure than the control side. Conclusion PDLLA plate containing rhBMP-2 has good biocompatibil ity and osteoinducibil ity to enhance fracture heal ing.