—In this paper we propose a multi-agent approach for traffic-light control. According to this approach, our system consists of agents and their world. In this context, the world consists of cars, road networks, traffic lights, etc. Each of these agents controls all traffic lights at one road junction by an observe-think-act cycle. That is, each agent repeatedly observes the current traffic condition surrounding its junction, and then uses this information to reason with condition-action rules to determine in what traffic condition how the agent can efficiently control the traffic flows at its junction, or collaborate with neighboring agents so that they can efficiently control the traffic flows, at their junctions, in such a way that would affect the traffic flows at its junction. This research demonstrates that a rather complicated problem of traffic-light control on a large road network can be solved elegantly by our rule-based multi-agent approach.