The importance of designated earthquake shelters has been emphasized in South Korea. However, the lack of an appropriate evaluation process for shelter candidates often leads to an ad-hoc selection of earthquake shelters, resulting in inefficient disaster management. To address this issue, this study proposes an evaluation framework based on multi-criteria decision-making to measure the suitability of earthquake shelters quantitatively. First, available indicators for shelter evaluation are extracted from the relevant literature, which are then streamlined and grouped using text mining to represent evaluation areas. Next, the selected indicators are standardized to the same value scale, and the Analytic Hierarchy Process (AHP) is used to derive the relative importance of each evaluation area to compute the weighted sum total score of each candidate. The developed evaluation framework is applied to a case study of earthquake shelter selection in the city of Songdo-dong. The results show that there are additional shelter candidates suitable for earthquake evacuation in addition to the current designated shelters. This study can serve as a basis for objective shelter selection that would improve the identification and utility of designated shelters.