Modern urbanization, while enhancing living standards, poses significant navigation challenges for visually impaired individuals. Existing assistive technologies, though improving, often lack crucial feedback integration and comprehensive environmental understanding. This paper proposes a novel assistive navigation system integrating real-time object detection, distance calculation, and directional guidance to address these limitations. Leveraging YOLOv5 for efficient object recognition, the system calculates distances based on object pixel height and a known focal length. Directional cues (left, right, stop) are provided to aid navigation. The system utilizes the COCO dataset for training and processes RGB images directly, optimizing performance. Performance analysis demonstrates the system’s ability to detect objects, estimate distances, and provide navigational assistance, though some distance variations occur due to environmental factors.