OBJECTIVE: To study the clinical efficacy of double bundle posterior cruciate ligament (PCL) reconstruction with remnant preservation. METHODS: From January 2007 to November 2011, 50 patients with PCL rupture met the inclusion criteria were divided into two groups: remnant preservation group (RP group) and remnant resection group (RR group). There were 19 males and 7 females in the RP group, ranging in age from 18 to 55 years, with a mean of (32.250 +/- 11.085) years old. The duration from injury to operation ranged from 2 to 66 months, with an average of (17.481 +/- 3.568) months. Among the RR group, 17 patients were male and 7 patients were female, ranging in age from 20 to 54 years old, with an average of (31.458 +/- 9.569) years. The duration from injury to operation ranged from 3 to 72 months, with a mean of (19.354 +/- 3.950) months. The patients in both groups suffered from instability of knee joint, got a positive result of posterior drawer test. In the RP group, the intercondylar notch remnant fiber, scar tissue and synovial were preserved in operation, only the free ligament in the intercondylar notch was resected. In the RR group, the remnant fiber, scar tissue and synovial tissue of adhesive parts were resected. In both groups, autologous semitendinosus and gracilis tendon double-bundle PCL reconstruction were carried out, the tibia was fixed with an absorbable interference screw with post-tie fixation, and the femur side was compositely fixed with absorbable interference screws and suspending fixation. Each patient received both subjective assessment (IKDC subjective evaluation, Lysholm scoring and Cincinnati rating) and objective clinical assessment (IKDC objective evaluation and Kneelax 3 tibia backward measurement) before operation and two years after operation. RESULTS: IKDC subjective evaluation: 92.167 +/- 4.177 in the RP group,which was higher than 87.542 +/- 5.687 in the RR group (P = 0.010). Lysholm scores: 90.917 +/- 4.413 in the RP group, which was higher than 87.083 +/- 5.149 in the RR group (P = 0.027). Cincinnati knee scores: 92.125 +/- 4.003 in the RP group, which was higher than 87.791 +/- 6.665 in the RR group (P = 0.027). IKDC objective evaluation:no significant statistical differences between RP group and RR group. Kneelax 3 assessment : tibia backward test with Kneelax 3 under 132 N showed no significant statistical difference between RP group and RR group, which were (3.958 +/- 0.693) mm and (4.029 +/- 0.846) mm respectively (P = 0.795). CONCLUSION: The study shows a significant advantage of remnant fiber preservation than remnant fiber resection in double-bundle PCL construction in terms of subjective knee function recovery after operation. There is no significant difference in postoperative knee stability.