The success of international construction businesses is highly subject to market uncertainties in association with host country attributes. That's because going international requires a large-scale investment and target country selection is a key component of corporate decision making. However, despite emerging interests and the need for systematic country evaluation, most existing studies have focused on specific regions or countries and have lacked comparative analyses among diverse countries using objective information. To address this limitation, this study suggests a data-driven approach to country classification using a two-stage model. The first stage represents a process of macro-classification to screen negative countries regarding country attractiveness. The second stage is a micro-classification procedure whereby promising countries are identified based on their level of past project performance. The model utilizes not only secondary data from internationally reputable institutions but also actual performance data of Korean firms from 1990 to 2014 to evaluate 32 countries. Using a factor analysis and two-step cluster analyses, the model allows us to cluster 32 countries by characteristics and explain the current status of each country. The proposed approach enables firms to quantitatively evaluate candidate countries and make better decisions at the early stage of corporate strategy development.