Data integration is becoming increasingly critical due to the vast amounts of information available in the Web and to the need for services that use information from different sources. Within the semantic Web, ontologies are crucial to provide data sharing and operability. However, when applications and services produced by different developers interact, we need to allow data to be shared and reused across distinct ontological frameworks. The process of establish “agreements” between different ontologies is called alignment, and is usually achieved by finding correspondences between their entities. In this paper we present an improvement of a fuzzy multi-layer architecture to perform ontology alignment. We use fuzzy logic techniques to combine different similarity measures among ontology entities, taking into account criteria such as the terminology, and the internal and relational structure of the concepts. This work was validated using the tests of the Ontology Alignment Evaluation Initiative (OAEI). The results show that the proposed techniques outperform previous approaches in terms of precision and recall.