Computational thinking (CT) is vital for success in numerous domains. However, the nature, definition, and scope of CT are ill-defined, and research on how best to develop CT is very limited. This study focused on how thinking styles and STEM attitudes have effects on computational thinking. Using a proportionate stratified random sampling procedure, 1195 students from two universities were surveyed. A structural equation modeling analysis showed that students' thinking styles and STEM attitudes directly predicted their computational thinking skills and that thinking styles mediated the relationship between STEM attitudes and computational thinking skills. Thinking styles and STEM attitudes are strong predictors of CT skills. Based on the results, we recommended that the conceptualization of CT be broadened to reflect its trans-disciplinary nature within the context of STEM education. This study adds to the limited theoretical understanding of CT and CT-predictors in higher education, which has been studied much less than in K-12 education.