Although traditional analytical techniques can characterize learners' internal cognitive abilities laterally and indirectly, it is difficult to present learners' learning development characteristics and changes comprehensively and dynamically. Epistemic network analysis, on the other hand, is an emerging educational research method that integrates qualitative and quantitative analyses and is able to visually and dynamically characterize the developmental trajectory of students' interactive dialogues. Therefore, this study used epistemic network analysis to dynamically track students' interactive dialogues. It was found that (1) there was a difference in the use of students' interactive dialogues between the intervention and control groups, with students in the intervention group using more exploratory dialogues, while students in the control group used more cumulative dialogues. (2) Observed from the time perspective, the development of interactive dialogues of the students in the control group was not significant, and the interactive dialogues of the students in the intervention group developed faster in the early stage and slower in the later stage. (3) The conversation patterns of the students in the excellent group focused on exploratory conversations and were mainly based on reasoning and reflection, and most of the students in the excellent group came from the intervention group. And vice versa for the control group.