Holonic Multi-Agent Systems (HMAS) are a convenient and relevant way to analyze, model and simulate complex and open systems. Accurately simulate in real-time complex systems, where a great number of entities interact, requires extensive computational resources and often distribution of the simulation over various computers. A possible solution to these issues is multilevel simulation. This kind of simulation aims at dynamically adapting the level of entities’ behaviors (microscopic, macroscopic) while being as faithful as possible to the simulated model. We propose a holonic organizational multilevel model for real-time simulation of complex systems by exploiting the hierarchical and distributed properties of the holarchies. To fully exploit this model, we estimate the deviation of simulation accuracy between two adjacent levels through physics-based indicators. These indicators will then allow us to dynamically determine the most suitable level for each entity in the application to maintain the best compromise between simulation accuracy and available resources. Finally a 3D real-time multilevel simulation of pedestrians is presented as well as a discussion of experimental results.