Abstract: s: Objective To study the outcome in extensor mechanism after total knee arthroplasty (TKA) and assess its patellar tracking method. Methods One hundred and sixty-nine patients (199 knees) including 75 males (80 knees) and 94 females (119 knees) who underwent TKA in our hospital from March 2008 to February 2012 were enrolled in this study. Of these patients, 6 (8 knees) were diagnosed with ankylosis knee due to rheumatoid arthritis, one (1 knee) was diagnosed with ankylosing spondylitis, and 162 (190 knees)were diagnosed with osteoarthritis. The patients were divided into patella replacement group (n= 40 with 46 knees) and non- patella replacement group (n=129 with 153 knees). Their average age was 55.2 years (36-83 years). The extensor mechanism was readjusted in both groups. The patella replacement was decided according to its indications and contraindications proposed by Burnet et al. Whether the outcome of patellar tracking was satisfactory was judged by "no thumb test" and "towel clamp test". Results The patients were followed up for 31 months (18-48 months). No significant difference was found in complications between patellar replacement group and nonpatellar replacement group(P> 0.05), which occurred in 2 patients (2 knees) of patellar replacement group (accounting for 4.35%)and in 4 patients (5 knees) of non- patellar replacement group (accounting for 3.27%). No significant difference was observed in the KSS (Knee Society Score) score between the two groups 3 years after operation (P> 0.05). Conclusion Adjustment of the extensor mechanism can optimize the patella tracking after TKA. Combined “no thumb test” and “towel clamp test” can accurately assess the patellar tracking. Patella replacement does not affect the clinical outcome 3 years after operation.