The patellar tendon is known as the strongest material for the substitution of anterior cruciate ligament. There are many technical difficulties in reconstruction of the ligament with this tendon since patellar aponeurosis is weak, and the length of the tendon is restricted, and the adquate insertion in proper place is difficult. Authors harvested the middle 1/3 of patellar tendon c bone block of tibial tubercle and patella with the superficial layer of the quadriceps tendon left attached. The large triangular tibial tubercle piece was impacted into the femoral tunnel from distally to proximally, of which the orifice is the very place of the middle of the original anterior cruciate ligament. Two bundles of 0-dexon sutures made at the tibial tubercle portion, were pulled through to different outlet of femoral condyle during this procedure and were tied over the cortical bone of lateral supracondylar region. The tendon twisted 180°(clockwise for the right knee and counter-clockwise for the left knee) to adapt to screwhome movement. The patellar piece was fixed at the isometric point in full flexion and extension, which is usually the middle point of original tibial insertion c one or two barbed staples made at the original position of tibial insertion of the ligament. We tried this methods in 26 knees (12 vascularized grafts and 14 free grafts) and was followed for more than one year and the results was compared with those of 12 knees of original McIntosh operation (follow up period: 16-72 months, average 34.3 months). The instability were evaluated c Lachman test and OSI sagittal knee tester. The knee of negative Lachman test & pivot shift test was defined as "normal" The results were as follows;l. In modified McIntosh operation group negative Lachman test was noted only in 6 knees (50.0 % ) and there were 2 knees (16.7% ) of GI, II & III, respectively. 2. In vascularized bone block patellar tendon graft group II out of 12 knees (91.7%) showed negative Lachman test and only one knee (8.3%) revealed G II. 3. In free bone block patellar tendon graft 13 out of 14 knees (92.9%) showed negative Lachman test and the other one knee (7.1%) revealed GI. 4, The overall success rate (normal) of bone block patellar tendon graft was 92.3%. The success rate of patellar tendon bone block operation was significantly higher than that of modified McIntosh operation (χ²=6.48, df=1, P=0.0109) and there was no statististical difference between vascularized and free graft group (χ²=0.39, df=l, P=0.5322).