In the Technological Institute of Orizaba, a project was carried out to collect historical data of the Institute, where a large amount of multimedia data was retrieved. At the end of the project, it was detected that increasing multimedia content made it difficult to manage this data optimally. This work addresses the aforementioned problem since a system for the historical data management of the Institute is designed, which uses horizontal fragmentation to optimize content-based queries. Experiments using a cost model demonstrate that the system reduces the execution cost of content-based queries.