Clinical results of minimally invasive mitral valve surgery were retrospectively reviewed, and two different surgical approaches were compared in this study. Between 1997 and 2004, a total of 86 patients with mitral valve disease underwent minimally invasive surgery at the Yonsei University Cardiovascular Center. Age of patients averaged 41.6 +/- 14.0 years and 69 patients were female. Surgical approach included low- sternal incisions with mini-sternotomy, and right parasternal or thoracotomy approach. Either direct aortic or femoral arterial and bicaval cannulations were used in all patients. Patients were divided into two groups according to the method of surgical approach (parasternal (P) vs low-sternal (L)), and the results were compared. Postoperative NYHA functional class improved to 1.1 +/- 0.4 in all patients (no significant statistical difference between two groups). Mean wound length (P: 9.21 +/- 1.10 vs L: 11.24 +/- 0.82 cm, p < 0.05), and mechanical ventilation time (P: 10.42 +/- 4.36 vs L: 12.90 +/- 5.00 min, p=0.04) was significantly shorter in parasternal group, and mean operation time(P:294.74 +/- 59.41 vs. L:259.31 +/- 54.36 min, p=0.03) was significantly shorter in low-sternal group. Mean cardiopulmonary bypass time, and aortic cross clamp time was also shorter in low-sternal group without statistical difference. There were 2 minor wound complications in all patients (p=NS), and no hospital death. Comparing the two different surgical approach of minimally invasive mitral valve surgery, parasternal approach is thought to be more beneficial in reducing postoperative scar, and intubation time.