To address the challenges of excessive model parameters and low detection accuracy in print circuit board (PCB) defect detection, this paper proposes a novel PCB defect detection model based on the improved YOLOv8 (You Only Look Once v8), named CMME-Yolo. Specifically, the proposed model leverages the sliding window mechanism and the receptive field synthesis mechanism into the original YOLOv8 backbone. Particularly, this scheme proposes a fine aware aggregation paradigm to enhance the recognition of small-scale defects in PCB images and boost the detailed representation in YOLOv8. After effectively capturing fine-grained features from PCB images in the backbone, subtle features are strengthened through the SPD-Conv (Space-to-Depth) module in the neck, thus fulfilling fusion through multi-stream features. Finally, to achieve adaptive aggregation of fine-grained perceptual features, these processed features are fed into the bridge rectifier of the feature pyramid system—the Parallel Multi-stream Fine Aware Aggregation module (PMSFA). Extensive experiments are conducted on the PCB defect dataset released by Peking University to validate the effectiveness of the proposed model. The experimental results show that compared with the original YOLOv8, our approach achieved the improvement in mean Average Precision (mAP)@0.5 and mAP@0.5:0.95 by 4.9% and 12.1%, respectively. Meanwhile, the number of model parameters is reduced by approximately 14% compared with the original YOLOv8.