Text is the carrier of language, and language is the carrier of cultural soft power, if you want a country’s soft power to be enhanced, it will certainly start from the dissemination of the native language. This paper constructs a complex social network J-SEVIR model for the dissemination of Japanese text information with the help of complex network theory combined with the information dissemination model using graph theory as the technical support. The data about Japanese text information on Sina Weibo is used as the research object, and the data analysis is carried out through the dimensions of model simulation, real data comparison, and information dissemination enhancement strategies. The study shows that the peak number of Japanese text message dissemination nodes is 1.987*107, which is 41.32% and 28.94% higher than the peak number of dissemination nodes in the traditional SEIR model and BCIR model, respectively, and the peak number of disseminators of the Japanese text message dissemination enhancement strategy designed by the J-SEVIR model can be up to 0.62, whereas the number of Japanese text message dissemination counterattackers is only 0.12. Therefore, the number of Japanese text message dissemination counterattackers is only 0.12. Therefore, the establishment of Japanese text information dissemination paths with the help of complex networks based on graph theory can be used to provide new research perspectives for optimizing the effect of Japanese text information dissemination.