Entity Alignment (EA) is a fundamental task for integrating heterogeneous Knowledge Graphs (KGs). Current methods primarily explore and leverage associations within entity relation embeddings across KGs. However, these approaches often fail to fully exploit the rich multimodal information associated with entities and the valuable semantic constraints embedded within ontologies (or ontological schemas). This paper proposes a novel Ontology-enhanced Multimodal Entity Alignment framework, named OMEA. Our method constructs comprehensive entity representations by synergistically fusing structural, attribute, visual, and semantic information. Building upon this, it jointly embeds entities with key ontological meta-information, including class hierarchies and entity membership relations. This explicitly utilizes class hierarchies and semantic relationships to enhance alignment consistency and prune erroneous alignments. Extensive experiments conducted on two distinct real-world benchmark datasets demonstrate that OMEA, integrating deep multimodal understanding with holistic ontology reasoning(i.e., leveraging class hierarchies to enforce semantic consistency), significantly improves entity alignment accuracy compared to existing state-of-the-art methods.