Traditional screen defect detection faces challenges like low accuracy, high omission rates, and high manual workload. This study proposes the YOLOv8-FCG algorithm to enhance LCD screen defect detection. By incorporating FasterNet-inspired optimizations, including the C2f-Faster module for improved feature extraction and reduced computation, YOLOv8-FCG efficiently handles defect detection. The inclusion of the Convolutional Block Attention Module (CBAM) in the Neck enhances detection accuracy for defects of different sizes, while the Generalized Intersection over Union (GIOU) loss function improves bounding box precision. Experimental results demonstrate that YOLOv8-FCG surpasses traditional YOLOv8 in both speed and accuracy. This optimized model provides greater efficiency and real-time adaptability for high-resolution defect detection in industrial settings.