Traffic is one of the most important aspects in human daily life because traffic affects smoothness of capital flows, logistics, and other community activities. Without appropriate traffic light control system, possibility of traffic congestion will be very high and honder people's life in urban areas. adaptive traffic light control system can be used to solve traffic congestions in an intersection because it can adaptively change the durations of green light each lane in an intersection depent on traffic density. The proposed adaptive traffic light control system can be used to solve traffic congestion in an intersection because it can adaptive vhange the durations of green light each lane in an intersection depend on traffic density.The proposed adaptive traffic light control system prototype uses beagleboard -xM, CCTV camera, and AVR microcontrollers. We use computer vision technique to obtain information on traffic density combining viola-lones methods with Kalman Filter method. To calculate traffic lighttime of each traffic light in intersection, we use Distributed Constraint Satisfaction Problem (DCSP) . From implementation and experiments results, we conclude that Beagleboard-xM can be used as main engine of adaptive traffic light control system with 91.735% average counting rate.