Objective To study repair of osteochondral defects by using composite of autologous BMSCs and chitosan/HAP (CS/HAP) bilayered scaffold in rabbits and its feasibil ity as osteochondral tissue engineering scaffolds. Methods CS/HAP bilayered scaffolds were produced with CS and HAP using a lyophil ization and sintering method. The pore size of the scaffold was observed by scanning electron microscopy (SEM). Anhydrous ethanol substitution method determined its porosity. BMSCs were isolated from bone marrow and cultured by general bone marrow methods. Both CD44 and CD45 on the BMSCs surface were detected with immunocytochemistry to identify BMSCs. Cell-scaffold complex was made with BMSCs as seed cells and CS/HAP bilayered scaffold as carrier by fibrin glue planting technique. The distribution ofBMSCs in CS/HAP scaffold was tested by SEM. The osteochondral defect (4 mm in diameter and 3 mm in height) model was made in the right knee joint of 36 Japanese white rabbits, which were randomly divided into 3 groups. Defects were repaired with CS/HAP and BMSCs composite ( group A, n=12) and with CS/HAP implants (group B, n=12); defects were not treated as a control (group C, n=12). Histological evaluation and gross observation were carried out at 6 weeks (n=6 in each group) and 12 weeks (n=6 in each group) postoperatively. Semi-quantitative histomorphological analysis was done to evaluate the repair cartilage tissue according to the modified Wakitani grading scale. Results CS/HAP bilayered scaffold possessed a porosity of 76.00% ± 5.01% and pore size of 200-400 μm (mean 300 μm ) in CS layer, and 72.00% ± 4.23% and 200-500 μm (mean 350 μm) in HAP layer, respectively. BMSCs formed colonies within 10-14 days. Immunocytochemistry results showed BMSCs had positive CD44 expression and negative CD45 expression. At 6 and 12 weeks after operation, gross and histological observation showed that the cartilage defects were fully filled with regenerated tissue, but bone defects were partially repaired in group A; the cartilage and bone defects were partially filled with regenerated tissue in group B and group C. The modified Wakitani grading scale were 5.17 ± 1.17 and 3.20 ± 0.75 in group A, 9.00 ± 0.63 and 6.00 ± 0.89 in group B, and 10.00 ± 0.89 and 9.60 ± 0.82 in group C at 6 weeks and 12 weeks postoperatively, respectively; showing significant differences between group A and groups B, C (P lt; 0.05). Conclusion The novel CS/HAP bilayered scaffold possesses porous structure and will possibly become a newbiomaterial of osteochondral tissue engineering.