Object detection remains a complex yet pivotal task in computer vision, with applications spanning diverse fields such as retail, autonomous driving, medical imaging, and security. Despite significant advancements in object detection techniques, challenges persist in adapting these methods to real-world scenarios, where conditions often deviate from those in controlled training environments. These discrepancies can substantially affect detection performance. To address this, we propose a novel Fusion Model that integrates predictions from multiple state-of-the-art object detection models, including YOLOv11, Detectron2, and SSD. The Fusion Model is specifically applied to the critical task of medical polyp detection using the Kvasir-SEG dataset. Experimental results demonstrate that our Fusion Model surpasses the individual performance of each component model, underscoring its potential to improve accuracy and robustness in real-world applications.