The Cybersecurity Knowledge Graph (CKG) represents an invaluable integrated resource designed to support critical functions, including vulnerability mining and defense against cyber threats. Integrating multiple knowledge sources becomes easier with the application of entity alignment, a promising strategy that transcends the boundaries between disparate cybersecurity knowledge bases. Despite this potential, the inherent sparsity and specialization of various CKGs have caused significant performance reductions in current entity alignment methodologies when employed for CKG entity alignment tasks. This paper introduces an effective and efficient entity alignment framework, named CyberEA. This framework utilizes similarity interaction and entity type constraints for an initial entity alignment, supplemented by logical rules for completing the knowledge graph. Subsequently, CyberEA generates entity embeddings from multiple perspectives-name, attribute, and structure. CyberEA implements a Graph Convolutional Network (GCN) to train the entity alignment model and adopts Least Squares Support Vector Machines (LS-SVM) to integrate these perspectives. Experimental validation on multi-type entity datasets reveals that CyberEA consistently surpasses other contemporary entity alignment methods in metrics such as Hits@n, Mean Reciprocal Rank (MRR), and Mean Rank (MR).