Traditional cognitive linguistics defines metaphor as a purely linguistic phenomenon, and Forceville and other representatives have extended conceptual metaphor to the multimodal platform. Flags of all countries mainly manipulate graphics, colors, text, concrete image, typography design for ingenious layouts and through juxtaposition of these different symbols to interpret profound meanings of individual flags, which constitutes a multimodal metaphor, with the source domain and the target domain represented by distinct modes. This study intends to, from the perspective of multimodal metaphor theory, amalgamate Conceptual Metaphor Theory and Conceptual Blending Theory, select some epitomized flag cases to analyze, in an attempt to grasp how the overall meaning of multimodal metaphor is established, and ultimately excavate extensive space for multimodal metaphor study.