Apple harvesting remains a labor-intensive and costly operation, particularly in high-value regions such as Jammu & Kashmir, India. This paper presents a vision-based detection and localization system for automated apple harvesting using a Quantum-Attention-Enhanced YOLOv11 architecture. The proposed model eliminates handcrafted segmentation by integrating quantum-inspired attention within the YOLOv11 backbone, enabling robust detection of single and clustered ripe apples under varying lighting and occlusion conditions. Using the red-apple subset of the Fruits-360 dataset, the system achieved a validation accuracy of 96.78%, mAP@0.5 of 0.914, and an average localization error of 4.37 pixels, demonstrating precise spatial estimation suitable for robotic gripping. Operating at 20 FPS, the model ensures real-time performance. The results confirm the effectiveness of quantum attention in enhancing feature learning and contextual understanding, paving the way for scalable, intelligent robotic fruit-picking solutions.