Traffic jam is a problem that is often faced by people especially those living in crowded cities, which is wasting time and money. Mainly, traffic light has been the common cause for traffic jam. Almost all traffic lights that caused traffic jam, work conventionally with timer. This ineffective method sometime needs Police's intervention to reduce traffic jam. Basically, Police rules traffic based on traffic density. Indeed, this paper proposes adaptive traffic lights to control its timing through traffic density calculation on road pattern. The traffic density is obtained by using image processing on several road pattern. Later, the traffic density is utilized to calculate timing of the traffic light. There is a server that collects the data and controls the traffic light operations in a crossroad. The entire algorithms, including the calculation of traffic density and timing of traffic light are validated through real road condition. The results show that traffic density detection system has good accuracy and timing of traffic light is successfully calculated.