This study explores the potential for deep learning among students participating in the Innovation and Entrepreneurship Education Program (IEEP). Drawing on ecosystem theory and expected value theory, this research investigates how teacher-student and peer interactions can influence students' willingness to engage in deep learning, as determined by their self-efficacy and emotional value expectations. The study examines the relationships between perceived teacher-student and peer interactions, self-efficacy, emotional value expectations, and deep learning. A sample of 265 students from a Chinese university participated in the study. Research tools were developed using exploratory factor analysis (EFA) and partial least squares structural equation modelling (PLS-SEM) to test the research hypotheses. The results indicate that perceived teacher-student and peer interactions significantly impact students' self-efficacy and emotional value expectations, which in turn, influence deep learning behaviour. Self-efficacy and emotional value expectations mediate the relationship between perceived teacher-student and peer interactions and deep learning. The findings suggest that micro ecosystems can influence individuals' intrinsic belief values, which can, in turn, affect their behaviour. The study highlights the significant impact activities promoting such interactions can have on enhancing students' deep learning, innovation, and creativity abilities.