IntroductionIn 2020, COVID-19 forced higher education institutions in many countries to turn to online distance learning. The trend of using online education has accelerated across the world. However, this change in the teaching mode has led to the decline of students' online learning quality and resulted in students being unable to do deep learning. Therefore, the current research, aimed at promoting deep learning in the online environment, constructed a theoretical model with learning self-efficacy and positive academic emotions as mediators, deep learning as the dependent variable, perceived TPACK support, peer support, technical usefulness, and ease of use as independent variables.MethodsThe theoretical model was verified by SPSS26.0 and smartPLS3.0, and to assess the measurement and structural models, the PLS approach to structural equation modeling (SEM) was performed.ResultsThe study found that (a) positive academic emotions play a mediating role between perceived TPACK support and deep learning, perceived peer support and deep learning, and perceived technology usefulness and ease of use and deep learning; (b) learning self-efficacy plays a mediating role between perceived TPACK support and deep learning, perceived peer support and deep learning, and perceived technology usefulness and ease of use and deep learning.DiscussionThe findings of this study fill the gaps in the research on the theoretical models of deep learning in the online environment and provide a theoretical basis for online teaching, learning quality, and practical improvement strategies.