Due to its major role in assuring services based on transportation, urban traffic is no more a facultative subject. The common step between all existing work related to this area is modeling urban traffic. Different approaches are available, and each one has benefits and drawbacks, but the main obstacle behind many of existing approaches is the availability of data to use in model building. In this paper we propose a dataless solution based on activities to model traffic demand of a generated synthetic population, and a microscopic multiagent system to simulate and analyze urban traffic in Moroccan capital Rabat.