Today, traffic management is crucial in smart cities to manage easy mobility, alleviate congestion, and support emergency response. The increasing prevalence of vehicles in numerous locations makes it challenging for the system to handle dynamic circumstances. Thus, an integrated technique is required to make the system adaptable and intelligent. This research proposes a Multi-Feature AI-Driven Traffic Management System that combines IoT and machine learning, emphasizing adaptive, collaborative, and emergency-aware decision-making. The utilization of Random Forest algorithm for real-time traffic forecasting to make the system autonomous. Along with adaptation, the proposed system is also collaborative, so different entities can coordinate to handle the traffic well without human supervision. A comparative study is conducted to evaluate the model’s performance against existing methods. In addition, a real-world application with the nearest traffic volume estimation added for increased accuracy and responsiveness. The results show the efficacy of our method in maximizing traffic flow and enhancing urban mobility.