This paper describes the development of a traffic controller using fuzzy logic, combined with the analysis of video sequences through machine vision techniques. The controller is able to automatically manage the traffic flow in a set of intersections, giving priority to the traffic lights of roads where there are emergency vehicles waiting. The system uses a classification algorithm, which is trained in order to detect any vehicle the scene and an image processing algorithm that identifies emergency vehicles within the previously detected vehicles. Using the information of video sequences acquired with CCD cameras installed on the intersections, the system choose the sequence of actions that improves the traffic flow, so as to increase the mobility in the road where the emergency vehicle is detected. Results show that the system is able to detect vehicles in real time. Also, the system adapts in an efficient and fast way to the changes in traffic flow in order to establish a priority road for emergency vehicles.