Visual characteristics are important factors in the transmission of images and animation for semantic interpretation; they have the functions of symbolic conversion and representation of the information content, and are also media between the viewer and the information in the image cognition process. The author had published the related researches for discussions of spreading on semantic network, and deeply investigated on image cognitions of Chinese poetry. This study applied content brainstorming narrative methods to engage in exploration of spreading activation model (SAM), and used the poetry of Wang Wei, who is known as the Buddha of Ancient Chinese poetry, Autumn Evening in the Mountains, to collect keywords for textual association and interpretation. This study also designed visual images that correspond to the associated keywords, based on dynamic visual focus of image content. Finally, this study selected 20 fourth grade students in Taiwan as the test subjects, and the experiment method was used to collect the described associations of students after they viewed the images, the lexical database formed by spreading vocabulary. The research findings aim to help explore interpretive relationship between the texts and images in poetry, providing a reference basis for visual transmission design.