In this work a unied approach for modeling of micro-robot swarms and for development of the controlling software architecture for individual robots is presented. The approach leads to swarm models with self-organized behavior and re ects the complexity of the considered swarm scenarios. The application of the presented approach is demonstrated by example scenarios with several complexity levels. Additionally, some techniques for the investigation of the phenomena of self-organization are discussed. These techniques allow for instance the determination of areas in the parameter space, i.e. the parameter settings leading to a specic aimed behavior. Through our simulation experiments we demonstrated, that the hardware requirements for applications using real micro-robots can be determined.