Traditional English teaching methods primarily involve the teacher explaining textual knowledge, often leading to inefficient use of classroom time. This study employs functional linguistics, visual grammar, discourse analysis, and interactive communication theories to analyze multimodal discourse in teacher-student interactions within college English classrooms. The study employs both quantitative and qualitative methods, using ELAN annotation software to manually annotate and analyze classroom teaching videos and lessons on a second-by-second basis across multiple modalities. The results indicate that modal distribution varies across different stages of English teaching, with verbal moderation prevailing and non-verbal moderation serving as a supplement. Verbal interactions show that teacher discourse constitutes 42% of total course time, 1.24 times more than student discourse. These findings offer data-driven insights for teachers to refine their strategies and boost interaction in college English classes.