abstract This paper gives an overview of the modelling and simulation system MIMOSE, which consists of a model description language and an experimental frame for the simulation of models. The main purpose of the MIMOSE project was the development of a modelling language which considers special demands of modelling in social science, especially the description of nonlinear, quantitative and qualitative relations, stochastic innuences, birth and death processes, as well as micro and multilevel models. At the same time, describing models in MIMOSE should not burden the modeler with a lot of programming and implementation details. Furthermore, the language concept should support the development of structured, homogeneous simulation models, which improves the transparence of the \model programming process" and makes model descriptions and even the corresponding simulation results easier to understand. To reach these goals the modelling concept of MIMOSE is based on ideas from general systems theory and the language structure is innuenced by paradigms of functional programming languages. Modelling and simulation with MIMOSE is demonstrated in detail by developing both a macro model and a multi-level model of the well known prey/predator model by Lotka and Volterra. Compared to the macro model, in which only the behaviour of the number of preys and the number of predators is deened by diierence equations, the multilevel model describes the individual behaviour of preys and predators on the micro level and its eeects on the macro level as well as feedback eeects from the macro to the micro level. Furthermore, single preys and predators can be born and can die, according to their individual birth or death probability. MIMOSE provides a powerfull modelling language combined with a user friendly experimental frame for simulating and analysing models. Therefore, this approach can be seen as one step towards the development of more general modelling and simulation tools in social science.