This study seeks to explore relations of visual-verbal modes and figure out application principles in China’s College English Classroom (CEC). It takes data from two files: (1) videos of two excellent CEC teachers; and (2) semi-structured interviews with them, within which it studies four modes in PPT or on blackboard presentation: image, words, dynamic and symbol. Three instruments—Multimodality annotation software ELAN, two-dimensional meaning-making tables and semi-structured interview are employed to facilitate both quantitative and qualitative analyses. The results showed features of frequency, timing and proportion of each mode summarized by ELAN and found out proper collocation of image, words, dynamic and symbol relies on intersemiotic relations, which are revealed as complementary and non-complementary. It further analyzed application principles of four modes under CEC context in China. Except principles of effectiveness, efficiency and appropriate collocation put forward by former studies, it complemented principle of modes’ transference to highlight the necessity to form students’ autonomy of visual-verbal modes.