Background The effect of prosthesis–patient mismatch (PPM) on clinical outcomes after aortic valve replacement remains controversial. We evaluated effect of PPM on long-term clinical outcomes after isolated aortic valve replacement in patients with predominant aortic stenosis. Methods We analyzed data from patients with predominant aortic stenosis who underwent isolated aortic valve replacement between January 1995 and July 2010. The indexed effective orifice area, obtained by dividing the in vivo effective orifice area by the patient’s body surface area, was used to define PPM as clinically nonsignificant (group I, 224 patients), mild (group II, 52 patients), moderate (group III, 39 patients), and severe (group IV, 36 patients). Results Early survival was not significantly different among the groups, but overall survival was decreased gradually in group IV. Overall survival at 12 years was lower in group IV than in group I (92.8% ± 2.7% vs 67.0 ± 10.1, respectively; P  = .001). Cardiac-related-death-free survival at 12 years was lower in patients with severe PPM. Left ventricular mass index decreased during the follow-up period in all groups. But left ventricular mass index was less decreased in group IV compared with groups I, II, and III. Age, severe PPM, and ejection fraction <40%, and New York Heart Association Functional Class IV were independent risk factors of overall survival on multivariate analysis. Severe PPM was an independent risk factor for cardiac-related death. Conclusions Severe PPM showed an adverse effect on long-term survival, and was an independent risk factor for cardiac-related death. In addition, patients with severe PPM showed less decreasing left ventricular mass index during follow-up.