The efficiency of the traffic flow is dependent on the traditional traffic signal control system. The traditional traffic signal control system follows the releasing of the traffic flow in interval times or fixed time schedules. That leads to heavy traffic flow and fuel consumption and also larger waiting times. Due to that people face many issues and also some risks will occur for the emergency services like ambulances and other service-related vehicles due to the heavy traffic in the urban areas. To overcome this problem, we implementing a solution called “Deep Reinforcement Learning–Based Adaptive Traffic Signal Control System for Real-Time Congestion Optimization” it offers an approach of using Deep Reinforcement learning that can help better use of the Deep-Q-Networks to bitterly use of the traffic management by predicting the Q-values for the different traffic signal actions it used to optimize the light changes under different traffic conditions. This type of simulation environment is organized by the SUMO (Simulation of Urban Mobility) for the network and traffic management and it uses the TraCI (Traffic Control Interface) it monitors the real time traffic data of continuous traffic flow. By collecting all these data we can compute and create a model that can manage the traffic signals, the model which can be built by the Deep Reinforcement learning Techniques and that model will be used to the better controlling of the traffic flow and reduces the waiting time for the vehicles and smooth traffic flow in the urban areas and helps to the people that they can reach their destination in early times and also helpful for the public transportation and reduce the chances of the road accidents and helps the environment by reducing the fuel consumption by minimal waiting times and also useful for services like ambulances and fire engines and etc in urban areas.