Information retrieval is one of the hot research topics in the field of computer science, and it has a wide range of application scenarios. However, it is mainly limited to specific domains or medium, so it’s unable meet the interaction needs of different user groups with different media. Aiming at the task of cross-domain information retrieval, this paper proposes a method based on a multimodal generative large model. Firstly, it analyzes the cross domain information retrieval technology in detail, summarizes the characteristics of information retrieval, and elaborates on the retrieval methods and processes. Then, a cross-domain information retrieval model based on multimodal generation is proposed. Finally, the effectiveness and reliability of the algorithm model are verified through experimental results. The results show that the proposed method achieves high accuracy in multimodal large models and achieves the best performance in cross-domain information retrieval tasks.