When infrastructures deteriorate, they are repaired to recover their states. Enormous records of infrastructure ' s repair have already accumulated. However, there is few research of the repair effect evaluation. In this paper, the authors conduct the statistical deterioration prediction, and propose the post evaluation method of repair effects. The mixed Markov deterioration hazard model and its hierarchical Bayesian estimation are employed as concrete model and estimation method, and the heterogeneity between deterioration processes before and after the repair is evaluated quantitatively. In addition, through the relative evaluation of deterioration rates, the post evaluation of repair effects is conducted. The relative evaluation is carried out by using Wilcoxon singed rank test between heterogeneity parameters before and after the repair, and repair effects can be expressed quantitatively as the test statistic. Furthermore, the proposed method is applied to the inspection data and the repair data of RC slabs to verify its effectiveness empirically.