This article describes an approach, called OPNets, for modeling real-time systems based on the object-oriented formalization of high-level Petri nets. To increase the maintainability and reusability of objects in Petri net modeling, the approach focuses on the decoupling of interobject communication knowledges and the separation of synchronization constraints from the internal structure of objects. To validate the overall system, which is composed of the hierarchically organized objects and interconnection relations, we used a two-step validation procedure that reduces the complexity and computational efforts required. As an illustration, a manufacturing cell with machining centers and robots is modeled using OPNets. The modeling experiences with OPNets demonstrate that the decoupling and separation of knowledges and constraints clearly enhances maintenance and reusability in real-time system modeling.