In this paper, we propose a reconciliation and cohabitation model of heterogeneous information in Decisional Information Systems DIS. Among this heterogeneous information are data that represent the analysis basis in DIS systems and metadata that represent a very important role to interpret the results and lead to a good analysis. In the current generation of DIS systems, we can find knowledge represented by ontologies, which is used to enjoy the semantic side offered by these systems. Our reconciliation model focuses on semantic mediation approach, based on the hybrid ontology architecture and semantic similarity measures to integrate each information type. Our modelling is based on a common model that represents a global view of heterogeneous information, allowing to find the mapping between different information types, to respond effectively to a posed query. To validate our work, we have applied our model on real business information, containing data, metadata and knowledge.