This paper first analyzes the characteristics and components of students’ core literacy in higher education institutions and constructs a flipped classroom teaching model under the deep learning perspective. Secondly, a Stacking integration-based model for student core literacy features classification is constructed by integrating Lasso regression, random forest, and XGBoost algorithms as base learners. Finally, a case study was conducted to verify the effectiveness of the flipped classroom model in cultivating students’ core literacy. The results showed that the p-values of moral quality, professionalism, humanistic literacy, and practical innovation of students under the flipped classroom teaching model were 0.022, 0.038, 0.012, and 0002, respectively, all of which were less than 0.05, with significant differences. The flipped classroom teaching model can effectively develop students’ core literacy, as shown by this.