Object detection plays a vital role in autonomous vehicles for safe and effective navigation through constantly changing environments. This study examines the combination of convolutional neural networks (CNNs) with transfer learning to enhance object detection capabilities in autonomous systems. Utilizing pre-trained models and customizing them to particular driving conditions, transfer learning minimizes the requirement for extensive labeled datasets while still ensuring high levels of accuracy. This proposed model aims to enhance the vehicle’s ability to detect pedestrians, vehicles, and obstacles under varying conditions such as lighting, weather, and traffic density. The results show that integrating transfer learning with CNNs achieves robust detection capabilities, making it a promising approach for real-time applications in autonomous driving systems.