Complex systems involve several business expertises that are designed as models in different modeling languages. These partial models are manipulated by different designers, and are thus generally heterogeneous (i.e conform to different metamodels). To create a complete view of the system, we proposed a process to organize partial models as a network of models through a virtual global model. As models evolve, changing elements involved in a correspondence, may cause the inconsistency of the global model. So, we have defined a process that automatically identify changes, classify them and treat their impacts on elements of other partial models in order to maintain the global model consistency.