The rapid development of the new media era has brought brand new opportunities for visual communication design, and the creation concept and expression of visual communication design have undergone obvious changes. In this paper, the ant colony algorithm is applied to automatically arrange graphic text in visual communication design to achieve the purpose of visual communication design innovation. Additionally, an image synthesis model based on generative adversarial networks is proposed for visual communication design, and examples are analyzed to analyze visual communication design in the context of future media. As a result of the study, the average loss values of the generator and discriminator of its generative adversarial network model are 9.894 and 2.319, respectively, and the Inception score (3.74) corresponding to the GAN is undoubtedly the highest, which has a high communication effect. This study enhances the effectiveness of visual communication design, promotes efficient communication and interaction with the audience, and promotes the innovative practice and theoretical exploration of visual communication design.