Abstract: British and American poetry has its obvious multimodality for its multi-sensory beauty of visual, auditory, olfactory and kinesthetic, which makes it possible for multimodal discourse analysis of American poetry aesthetic criticism. This paper constructs a framework of multimodal analysis of American poetry discourse and conducts a multi-modal discourse analysis and interpretation of poetry's context, the functional meaning of the poetry text, multimodal features inside and outside poetry text, and modal relation from context layer, content layer and expression layer of the poetic discourse, with Edgar Ellen Poe's To Helen as an example. Multimodal discourse study of poetry is an extensive exploration from the perspective of interdisciplinary research of British and American poetry. This research paradigm is expected to contribute to the British and American poetry criticism.