From the perspective of multimodal corpus, this study deeply analyzes the characteristics of the use of Chinese vocabulary and the law of pragmatic evolution in academic language writing scenes. The case analysis method is used to select academic texts and reports in the fields of science and technology, humanities and social sciences, medicine and other disciplines as the real corpus, and the comparative research method is used to compare the use of cultural vocabulary in different periods and disciplines. In the exploration of cultural vocabulary and multimodal collaborative application mode, three modes of complementarity, reinforcement and expansion were found. In the complementary mode, cultural vocabulary makes up for the shortcomings of other modalities in conveying abstract cultural connotations. The reinforcement mode is manifested as a multimodal enhancement of the ideographic effect of cultural vocabulary. The expansion mode is a multimodal extension of the semantic scope and application scenarios of cultural vocabulary. These patterns are influenced by factors such as cultural background, disciplinary specificity, and authors' writing intentions. Based on the theory of semantic evolution and pragmatics, this paper explores the pragmatic evolution of cultural vocabulary in academic writing, which has the characteristics of gradual and staged. Social development, subject development, and language exposure are the main drivers. Multimodal corpora provides abundant data for the study of cultural vocabulary, but there are challenges such as high labor cost, difficult accuracy control and high technical requirements for data annotation. This study provides enlightenment for improving the quality of academic language writing and promoting the development of cultural vocabulary research.