The surge in computer vision applications has amplified the need for precise and resilient car classification systems. Addressing this, our research delves into enhancing the accuracy and adaptability of systems classifying four car types: Audi, Mahindra Scorpio, Swift, and Tata Safari. Our approach used a rich training dataset of carefully curated and annotated images representing the targeted car categories, coupled with a validation dataset for model refinement. Utilizing the YOLOv8 deep learning model, we embarked on rigorous training and validation phases. A key emphasis was on combating overfitting, ensuring our model's wider applicability. Our efforts culminated in a commendable 91 % accuracy. This paper offers a gamut of insights: illustrative charts, detailed outcomes, and a thorough dissection of the confusion matrix. Our results highlight YOLOv8's prowess in car classification and its promise in bolstering object recognition systems, all while adeptly navigating overfitting challenges.