Seals are one of the most important elements of historical documents, as they perform an essential function of authentication, identification of the issuing authority, and legal and symbolic contextualization. Their traditional analysis, based on sphragistics, requires specialized skills and careful manual analysis, which is difficult to scale with large digitized archives. In recent years, machine learning techniques have demonstrated considerable potential in the automatic analysis of historical images; however, purely data-driven approaches are often inadequate for capturing the semantic and historical complexity of seals. This paper proposes a study on the representation of seals and their identifying characteristics using semantic web technologies. The aim is to develop a shared and universal semantic representation that can be populated with the information content of seals found in historical archives. This allows for their use, classified according to their properties, by creating virtual archives that can be consulted using recent and widespread virtual reality and immersive technologies.