Recent years have seen considerable growth of online multimedia databases, largely due to digitization processes in different sectors of society. Knowledge organization and representation strategies were used to qualify and enrich data and metadata from different types of documents and ensure persistent and interoperable online information structures. This study aimed to propose an ontological reference model to systematically organize metadata that describes multimedia documents based on different contexts and needs. The proposed model was based on the NeOn methodology and aimed to encompass the functional and nonfunctional requirements for the construction and reuse of ontology classes obtained by merging and aligning previously analyzed multimedia ontologies. This resulted in a comprehensive conceptualization to organize multimedia metadata for application contexts that deal with the semantic annotation of information entities produced and consumed in the web of data (Semantic Web). We concluded that advances in developing conceptual reference models for representing multimedia documents are the result of interdisciplinary efforts that drive progress in the production and use of more consistent and coherent metadata aimed at facilitating the cross-referencing, interconnection and aggregation of online information sources.