Multimedia discourse refers to the discourse that produces the overall meaning through the interaction of two or more communication modes such as image, language and sound. With the rapid development of multimodal discourse analysis in China, it has gradually become a research hotspot. This paper takes the Disney live action movie Mulan as the research object to explore the collaborative relationship between images and texts, to expand the research field of multimodal discourse analysis, and to verify its practicability and feasibility in the study of film discourse.