Background: The aim of this study is to evaluate our institutional results of the aortic valve replacement through minimally invasive approaches compared with conventional sternotomy. Materials and Methods: From August 1997 to July 2010, 838 patients underwent primary isolated aortic valve replacement. Of them, 73 patients underwent surgery through minimally invasive approaches (MIAS group) whereas 765 patients underwent surgery through the conventional sternotomy (CONV group). Clinical outcomes were compared using a propensity score matching design. Results: Propensity score matching yielded 73 pairs of patients in which there were no significant differences in baseline profiles between the two groups. Patients in the MIAS group had longer aortic cross clamp than those in the CONV group ($74.9{pm}27.9$ vs.. $66.2{pm}27.3$, p=0.058). In the MIAS group, conversion to full sternotomy was needed in 2 patients (2.7%). There were no significant differences in the rates of low cardiac output syndrome (4 vs. 8, p=0.37), reoperation due to bleeding (7 vs. 6, p=0.77), wound infection (2 vs. 4, p=0.68), or requirements for dialysis (2 vs. 1, p=0.55) between the two groups. Postoperative pain was significantly less in the MIAS group than the conventional group (pain score, $3.79{pm}1.67$ vs. $4.32{pm}1.56$; p=0.04). Conclusion: Both minimally invasive approaches and conventional sternotomy had comparable early clinical outcomes in patients undergoing primary isolated aortic valve replacement. Minimally invasive approaches significantly decrease postoperative pain.