Abstract: 【Purpose/significance】Analyzing the influencing factors of continuance usage behavior of MOOCs can deeply
understand learners' motivation and participation behavior so as to provide more decision support for managers.【Method/
process】A model is constructed to capture the dynamics of continuance usage of MOOCs, referring to Hidden Markov Mod⁃
el(HMM) based on the learning process framework of Illeris. Then data from Shanghai Learning Platform is collected, and
the Maximum Likelihood method is used to estimate parameters. Furthermore, the influence of website, course, peer effect
and teacher-student interaction on continuance usage of MOOCs was analyzed in different ability states, and the evolution
of learners' ability state was also explored.【Result/conclusion】Peer effect has positive effect on MOOCs usage of learners
with different abilities. High ability learners are also more concerned about the website safety and information related with
the course, whereas low ability learners are more likely to attend teacher-student interaction activities. There are four typi⁃
cal ability evolution roadmaps, the high ability learners have the accumulation advantage and the positive MOOCs continu⁃
ance usage behavior track, and are more likely to keep the high ability state.