<p>Product recommendation system is essential for improving user experience and helps in growth, especially in e-commerce and online platforms. These systems generate personalized recommendations by analyzing user behavior, preferences, and historical data to enhance customer satisfaction. This study presents a comprehensive review of deep learning-based recommendation systems, which have outperformed traditional algorithms in terms of scalability and accuracy. The deep learning technique includes Deep Neural Network, Convolutional Neural Network, Recurrent Neural Network, Autoencoders, and Neural Collaborative Filtering, highlighting their architectures with strengths and weaknesses across diverse domains. To achieve this, we comprehensively reviewed various state-of-the-art approaches from 2018 to 2025 for a product recommender system using deep learning techniques. There are 58 primary studies which has been taken from highly reputed journals indexed by SCIE, Scopus, and IEEE conferences. Also, a taxonomy is presented, which provides the overall overview of the recommender system. Lastly, the findings contribute a various open issue challenges which still exist in the recommendation system and also provide future directions based on the study of the literature review. The study will help the researcher in finding current trends in recommender systems using various deep learning techniques.</p>