Product recommendation systems are critical in improving the user experience, retaining customers and boosting sales on a wide range of digital platforms. Of the many available methods, Market Basket Analysis (MBA) presents a different angle on the problem by discovering patterns in past transaction data to uncover item associations. In this paper, we conduct a comprehensive survey of the developments, approaches and applications of MBA for product recommendation. We first review classical algorithms such as the Apriori, Eclat, and FP-growth, and, then, we analyze some recent extensions, such as temporal, utilitybased, and hybrid models that combine MBA with collaborative filtering and machine learning. We further provide examples of MBA's utility in retail, e-commerce, and personalized marketing, and tackle important issues, including scalability, sparsity, and dynamic customer behaviors. Finally, we point out a few promising research directions, such as the incorporation of real-time analytics, deep learning technology and context-aware recommendations into existing systems. This survey is intended to provide a foundational reference source for scholars and industry workers who want to understand and promote the MBA in intelligent product recommendation systems.