The current research deals with the development of an Activity Based Multi Agent System able to analyse and simulate the complex dynamics of sustainable mobility as derived from millions of choices performed by the individuals belonged to the system itself. With the objective to perform a dynamic decision support system for the phenomena comprehension at urban scale, the population sample questionnaire structuring, the “temporal” geodatabase construction and the agents behaviural rules extraction, via two different artificial intelligence techniques, represent the most relevant innovations experimented till now.