Agent based modelling is a paradigm that is widely used to simulate and research systems where each entity is autonomous and behaves independently. Currently, there are many diverse agent based models, however, aside from general guidelines there are no particular analysis of how such models should be designed and implemented. On the other hand, any agent based model can be broken into single agent and system level, abstraction levels that are used in the design of multi agent systems, moreover, agent based models are often implemented using mechanisms and tools from multi-agent systems. Author proposes that one of the ways this problem could be solved is classification of agent based models from the perspective of multi-agent systems which is then used to provide decision support for design of agent based models.