Accurate turn movement counts are critical for optimizing traffic flow and improving intersection management. This entry introduces an AI-powered system that automates turn-counting using video feeds from road intersections. By combining deep learning-based vehicle detection with real-time multi-object tracking, the framework extracts precise vehicle trajectories and classifies turn movements based on key zone crossings. To enhance accuracy, a smart filtering mechanism eliminates duplicate counts by considering only the first inbound and last outbound crossings per vehicle. Rigorous testing across diverse traffic conditions confirms the system’s reliability, even in high-density scenarios—offering a scalable, cost-effective solution for smart city traffic analytics.