To study the effect of the repair of rabbit articular cartilage defects by the composite of chondrogenic induction of autologous MSCs and autologous “two-phase” bone matrix gelatin (BMG). Methods Twentyfour healthy adult New Zealand rabbits weighing 2 to 3 kg were divided into group A, B and C with 8 in each. Autologous MSCsderived from group A were cultured in vitro and observed under inverted phase contrast microscope when enough cells through trypsinization transferring in vitro were obtained. Then the growth curves of 1, 3 and 5 passage culture of MSCs were drawn. The 3rd passage MSCs were induced into chondrogenic differentiation by adding TGF-β1 (10 ng/mL), IGF-1 (10 ng/mL) and vitamin C (50 ng/mL) in vitro. At 8 days after induction, the features of chondrocytes were observed under inverted phase contrast microscope, and immunohistochemical staining and Mallory staining were made. Getting out part of the il ium of group A and B, according to the method of Urist, the “two-phase” BMG was acquired. Chondrogenic induction of autologous MSCs was inoculated into the corresponding BMG to set up a composite of cell-carrier, and then it was observed through scanning electric microscope after 3 days of culture. The model of articular cartilage defects of rabbits was made: in group A, autologous cell-carriers were implanted; in group B, there only existed autologous BMG; in group C, there was nothing. At 8, 12 weeks after operation, the gross, HE staining and immunohistochemical staining were made, and grading scales were evaluated according to Wakitani histological grading method. Results Features of MSCs were as follows: the shape of primary cells was shotspindled and of passage cells was long. As to the growth curves of 1, 3 and 5 passage culture of MSCs, passage cells grew slowly for 3 days after being passaged and went into log-growth during the 3rd and the 7th days and into plateau later, but the 3rd passage cells grew best. Observation of MSCs after chondrogenic induction was performed: the shape of cells was ell iptical and the effect of induction was verified by the positive results of collagen type II, S-100 and Mallory staining. Under scanning electricmicroscope, the structure of BMG was good and cells were observed growing in it well. As far as repair of articular cartilage defects are concerned at 8, 12 weeks after transplantation, the defects in group A were repaired by the hyl ine-l ike tissue and the structures of the cartilage surface and normal cartilage were in integrity, and immunohistochemical staining of collagen type II was positive, while those in group B and C were repaired by the fibrous-l ike tissues and the surfaces were irregular. In Wakitanni histological score, at 8 weeks after operation, group A was (3.50 ± 1.51) points, group B was (10.00 ± 1.41) points and group C was (12.00 ± 0.93) points; at 12 weeks, group A was (1.13 ± 0.99) points, group B was (8.38±1.30) points, and group C was (10.13 ± 1.64) points. At different time points, group A was significantly better than group B and C, showing significant differences (P lt; 0.05). Conclusion Induced autologous MSCs and the composite with autologous “two-phase” BMG have the function to repair articular cartilage defects, and they are better than autologous BMG transplanted only or nothing transplanted.