Printed Circuit Boards (PCBs) play a crucial role in modernelectronic systems, necessitating robust quality assurance measures to ensure their reliability and functionality. Detecting and localizing defects on PCBs are vital tasks in this regard, with determining defect type being particularly important for effective troubleshooting and repair processes. Classical manual inspection methods are not efficient at all due to slow process and chances human error. This has led to the exploration of machine learning (ML) and Deep Learning (DL) techniques for automating defect detection and classification tasks. This paper presents a comprehensive review of literature on PCB defect detection using DL approaches. We synthesize existing methodologies, challenges, and opportunities in this domain. By focusing on recent advancements in DL models, this review aims to guide researchers and practitioners in adopting the latest and most effective techniques for PCB defects prediction. This paper focus on suggesting adoption of newer models, can potentially improve the accuracy and robustness of defect detection and classification in PCB quality control processes. Few possible modifications are also suggested to make existing DL models more efficient.