Urban traffic congestion has become a major challenge, causing long delays, higher fuel consumption, and rising pollution levels. Traditional traffic signal systems often rely on fixed timing cycles, which fail to respond to real-time traffic conditions and lead to inefficiencies. This paper presents a Smart Traffic Management System designed to address these issues by using real-time data to dynamically control traffic signals. The system uses cameras and sensors to monitor vehicle density on each road segment, automatically adjusting green light durations based on traffic volume. Roads with heavier traffic receive longer green signals, while less crowded lanes are given shorter ones—ensuring optimal use of intersection time. In addition to traffic optimization, the system includes number plate recognition to detect and record red-light violations, improving law enforcement. A dedicated pedestrian detection module further enhances safety by recognizing people waiting to cross and extending green signals when needed. To make the system even more efficient, machine learning and deep learning models are employed to predict traffic trends using both real-time and historical data, allowing the system to make proactive adjustments. Overall, this intelligent and adaptive approach aims to create a safer, smoother, and more sustainable urban traffic environment.