Learning behavioral engagement is an important learning feature that affects students’ learning achievement and persistence. As the main driving force of students’ actual performance, there is not effective evaluation model of learning behavioral engagement for MOOCs platform. It is helpful to observe the change rule of students’ participation and activity by analyzing the data of learning behavioral engagement. In this paper, the online study data and feedback learning behavioral engagement data from four MOOCs course were analyzed, and the predictive effect of behavior variables on learning behavioral engagement were studied. According to the results of data analysis, persistence, reflection, initiative and concentration are identified as the assessment dimensions of level which can represent learning behavioral engagement. Based on this, the evaluation model based on MOOCs platform is constructed. The establishment of the model can lay the foundation for the design of related evaluation and analysis tools.