Cardiovascular diseases (CVDs) are one of the major causes of death worldwide, and thus early and accurate detection methods are required. Recent advancements in deep learning, especially YOLO-based models, have shown excellent potential in medical imaging for automated CVD detection. This paper compares the performance of YOLOv8, YOLOv9, and YOLOv11 in detecting cardiovascular anomalies using key performance metrics, including F1-Score, Precision-Recall (PR) Curve, Precision (P-Curve), and Recall (R-Curve). Results with the YOLOv8 model showed an F1-Score of 0.93, PR-Curve value of 0.957, P-Curve value of 0.99, and R-Curve value of 0.98. The figures depicted the high accuracy and robustness of the models to identify cardiovascular conditions from medical images. Comparison between YOLOv9 and YOLOv11 depicts what improvements are available in these models and which shortcomings might be faced with each of the models, so its applicability could be discussed within the real world in medical diagnostics. In conclusion, the above study explains that YOLOv8 provides a valid solution method for CVD detection with high precision and recall. Future work would include the optimization of YOLOv8 by including more sophisticated transfer learning techniques and larger, more diverse datasets to further improve model generalization to the clinic.