The Program for International Student Assessment (PISA) aims to provide comparative data on 15-year-olds' academic performance and well-being. The purpose of the current study is to explore and compare the variables that predict the reading literacy and life satisfaction of U.S. and South Korean students. The random forest algorithm, which is a machine learning approach, was applied to PISA 2018 data (4,677 U.S. students and 6,650 South Korean students) to explore and select the key variables among 305 variables that predict reading literacy and life satisfaction. In each random forest analysis, one for the U.S. and another for South Korea, 23 variables were derived as key variables in students' reading literacy. In addition, 23 variables in the U.S. and 26 variables in South Korea were derived as important variables for students' life satisfaction. The multilevel analysis revealed that various student-, teacher-or school-related key variables derived from the random forest were statistically related to either U.S. and/or South Korean students' reading literacy and/or life satisfaction. The current study proposes to use a machine learning approach to examine international large-scale data for an international comparison. The implications of the current study and suggestions for future research are discussed.