Large-scale open courses have become a new way of online learning. Exploring the topology of the learner-course interaction network has an important role in improving learner participation and the activity of the MOOC platform. The study randomly crawls the learner data of "China University MOOC", uses complex network analysis tools to explore the evolution of the degree distribution of learners, and based on the learning behavior modeling of the learners, proposes a learner learning gain output model, and uses Pajek software to The two-mode network is mapped to the learner's one-mode relationship network; at the same time, a regression model is constructed by combining the attribute data of the learners with each component of the attribute data that influences the learner's participation in the course.