Objective To evaluate the influence of patellar replacement on total knee arthroplasty by comparing with non pattelar replacement. Methods Between September 2010 and November 2010, 63 patients (63 knees) with osteoarthritis who met the selection criteria and underwent total knee arthroplasty, were randomly divided into 2 groups: patellar replacement in 32 cases (replacement group), non patellar replacement in 31 cases (non pattelar replacement group). There was no significant difference in gender, age, disease duration, osteoarthritis grading, the clinical and functional scores of American Knee Society Score (KSS), the patellar tilt angle, tibiofemoral angle, and patellar ligament ratio between 2 groups (P gt; 0.05), they were comparable. After 6 weeks, 3, 6, and 12 months of operation, clinical and imaging evaluation methods were used to assessment the effectiveness. Results Primary healing of incision was obtained in all patients of 2 groups. Deep venous thrombosis occurred in 6 cases of replacement group and in 8 cases of non pattelar replacement group. All patients were followed up 12 months. The postoperative incidence of anterior knee pain in replacement group was significantly lower than that in non pattelar replacement group (P lt; 0.05) at 3, 6, and 12 months after operation. No significant difference was found in the postoperative KSS clinical score between 2 groups at each time point (P gt; 0.05). The joint function score of the replacement group was significantly higher than that of the non pattelar replacement group at the other time point (P lt; 0.05) except the score at 6 weeks and 3 months. Significant difference was found in the patella score between 2 groups at 12 months (P lt; 0.05), but no significant difference at the other time points (P gt; 0.05). X-ray film showed no patellar fracture and dislocation, or loosening and breakage of internal fixation. At 12 months after operation, the tibiofemoral angle, the patellar ligament ratio, and the patellar tilt angle showed no significant difference between 2 groups (P gt; 0.05). Conclusion Patella replacement can improve knee function score and the patella score, and reduce the incidence of postoperative anterior knee pain.