Multimodal discourse analysis reveals the dynamic interplay between different modalities, transcending the traditional language-centric approach of single-modal discourse analysis. Using multimodal discourse analysis, the study draws on the translation practice of dialogues in The Butterfly Lovers and clarifies the roles of and relationship between different modalities (i.e., verbal and non-verbal) in constructing meanings in audiovisual translation. The study shows that the relationship between different modalities is complex and dynamic, with one modality potentially complementing, reinforcing or interrupting another in conveying information. Hence, in addition to pure linguistic conversion, translating Chinese drama into English also entails harmonious coordination of visual and auditory modalities to captivate the audience's senses. Furthermore, translators should leverage the multimodal way of presenting messages and resolving translation problems more tactfully, so as to achieve optimal translation and cross-culture communication outcomes. Theoretically, this research contributes to the understanding of multimodal aspects in dialogue translation by describing specific translation phenomena, thus expanding research perspectives. On a practical level, it seeks to enhance translation quality from a multimodal standpoint, fostering the appreciation of Chinese drama in foreign lands.