The invention relates to an indocyanine green fluorescence combined multimodal image-guided liver portal vein watershed developing method, and belongs to the technical field of medical images. According to the method, artificial intelligence technologies such as virtual reality, augmented reality and the like are integrated with traditional intraoperative ultrasound and indocyanine green fluorescent staining technologies, so that a surgeon can visualize three-dimensionally complex liver anatomical structures, accurately locate tumors in an operation, plan an operation scheme, accurately achieve developing of a target portal vein watershed, monitor and correct a liver parenchyma disconnection plane in real time, and perform accurate anatomical liver resection. In a liver operation, the method can display a portal vein branch blood flow supply area for supplying tumors in real time, accurately guide a liver resection range, can not only remove micrometastatic tumors that may exist in theportal vein, but also can ensure blood supply of the remaining liver, improve accuracy and safety of the operation, reduce operative complications, and improve efficacy of tumor operation.