Objective To explore the preparing methods in vitro and test the cl inical appl icabil ity of implantation in vivo of bone marrow stromal stem cells (BMSCs)-biphasic scaffold to repair defects of cartilage and subchondral bone and tocompare the differences in repaired outcomes of composite, single biphasic scaffold and rabbits themselves. Methods The upper chondral phase and the lower osseous phase of the plugs, using poly-lactic-co-glycol ic acid (PLGA), hydroxyapatite (HA), and other biomaterials, were fused into carrier scaffold, on which collagen type I (Col I) was coated. The surface and inner structure of bi phasic scaffold were observed under scanning electron microscope (SEM). BMSCs was isolated from the bone marrow of tibia and femurs of young New Zealand rabbits using centrifuging and washing, and their morphologies and adherences were observed everyday. Then BMSCs were inoculated on the surface of scaffold to form BMSCs-scaffold composites. Osteochondral defects were surgically created on articular surface of femoral intercondyles of 30 New Zealand rabbits, which were divided into groups A, B and C. In group A, a bi phasic osteochondral composite were implanted into defect, BMSCs and biphasic cyl indrical porous plug of PLGA-HA-Col I in group B, and group C was used as a control without implant. Specimens were harvested to make macroscopic and histological observations at the 1st, 3rd, 6th, and 9th months after operation respectively; meanwhile immunohistological and micro-computed tomography (micro CT) examinations were performed and graded at the 9th month after operation. Results SEM showed an excellent connection of holes in the biphasic scaffold infiltrated by Col I. Optical microscopy and SEM showed a good growth of BMSCs in scaffold without obvious cellular morphological changes and an accumulation in the holes. Macroscopic samples showed a resistant existence of defects of group C within 9 months; the scaffold completely degenerated and chondral-l ike tissue formed on articular surface with partly collapses and irregular defects in group A; and smoother surface without collapses and approach to normal with texture of new regeneration in group B. There were statistically significant differences in macroscopic results (P lt; 0.001), group B was superior to group A, and group C was the worst. The micro CT showed good repairs and reconstruction of subchondral bone, with a acceptable integration with newborn chondral-l ike tissue and host bone in group B. Quantificational analysis of relevantparameters showed no significant differences. Histological results showed inflammations located in defects at the 1st month, new tissue grew into scaffold at the 3rd month; new chondral-l ike tissue crept on the margin of defects and biphasic scaffold degenerated completely at the 6th month, and lots of collagen formed in subchondral bone with major fibrocartilage on chondralarea at the 9th month after surgery in groups A and B. In groups A and B, immunohistological observations were weak positive for Col II and positive for Col I. Conclusion Biphasic scaffold implanted in body can induce and accelerate repair of defects of articular cartilages which are mainly filled with fibrocartilage, especially for subchondral bone. Scaffold combined with BMSCs has the best repairing effects 9 months after implantation.