This research paper delves into a comprehensive investigation of defect detection methods in printed circuit boards (PCBs) by employing the YOLOv5 architecture. As the complexity of PCB designs continues to escalate, there is a growing need for robust defect detection techniques to ensure the reliability and quality of these essential electronic components. The core contribution of this work revolves around the identification and localization of singular PCB defects using the YOLOv5 architecture. Furthermore, the research extends its scope to explore the capabilities of YOLOv5 in simultaneously detecting two distinct PCB defects. The experiments conducted are based on publicly available PCB defect datasets and involve various preprocessing steps to enhance the accuracy of defect detection. To assess the effectiveness of the proposed approach, the evaluation process relies on several key metrics. Mean Average Precision (mAP), precision–recall (P–R) curves, and confusion matrices play pivotal roles in quantifying the model’s performance. The outcomes of this evaluation reveal that the YOLOv5 architecture demonstrates notable success in detecting specific PCB defects, particularly shorts and missing holes. These findings are substantiated by higher MAP scores, underlining the model’s efficacy in addressing these critical issues within PCBs.