After the outbreak of COVID-19, the literary and artistic works about COVID-19 have appeared one after another. As a new branch of discourse analysis, the Multimodal Discourse Analysis has always been an important way for studying multimedia discourses. Based on the theory of visual grammar, this study intercepts and makes a qualitative analysis of the images with research significance in the documentary Together Against COVID-19 produced by CGTN. In this way, this study has drawn a conclusion that the documentary editors use picture language to create a responsible, friendly, and also dignified national image of China, to describe the emotional state of the masses in China in the process of COVID-19, and to achieve the purpose of conveying emotions and shaping positive national image through different aspects.