The education landscape has been radically transformed after the COVID-19 pandemic. Therefore, this research aims to devise innovation and apply artificial intelligence technology to novel learning styles. There are three objectives for developing innovations in this research: to study the context of the hybrid learning model to promote learning achievement for university students, to construct text-mining predictive models to recommend appropriate learning styles for students, and to study the efficiency of the text-mining prediction models. The target population and experimental group were 63 students from the School of Information and Communication Technology at the University of Phayao. The research experiment was controlled, and students could freely choose their studies. The research activities consisted of two domains: online and on-site learning management. Research tools include fifteen learning activities, questionnaires, and text-mining predictive models. The research found that learning styles have similar influences on learning achievement, learners clearly have the same acceptance and attitude toward different learning styles, and text-mining predictive models can efficiently classify and present appropriate learning styles. Most students choose to study online, but the learning achievement level of students who use the on-site system is higher than that of those who selected online learning. Finally, the predictive model using the text mining technique is highly efficient and accurate, with an accuracy of 98.75 (±3.95)%. Therefore, the research has significant implications for use, dissemination, and as a model for improving the quality of education at the higher education level.