This paper proposes a novel solution called Multi-Modal Entity Alignment (MMEA) to address the problem of entity alignment in a multi-modal view and designs a novel multi- modal knowledge embedding method to generate the entity representations of relational, visual and numerical knowledge.