A prospective series of 114 consecutive minimally invasive surgeries for total knee arthroplasty was performed using the quadriceps-sparing approach at the beginning. Intraoperatively, when the knee was in 45 degrees to 60 degrees of flexion, lateral patella subluxation was evaluated. A progressive quadriceps tendon incision with a 1-cm increment was applied if the patella could not be completely slid. The mean follow-up time was 24 months. There were 3 groups according to the length of quadriceps incision: group A (17 knees) had no or 1-cm quadriceps incision; group B (60 knees) had 2-cm incision; and group C (37 knees) had 3-cm incision. The average operative time, blood loss, pain score, preoperative range of motion and postoperative range of motion at 2 weeks, 6 weeks, 12 weeks, and 3 months were not significantly different among groups. Patient ability for early ambulation (sitting, knee straightening, standing, and walking) was indifferent between groups A and B; however, this was significantly delayed in group C. In conclusion, minimally invasive surgery for total knee arthroplasty with 2-cm quadriceps incision or strict quadriceps-sparing approach provided no difference on early ambulation.