Scour analysis stands as a pivotal factor in guaranteeing the safety and soundness of hydraulic structures, such as bridge piers and abutments, which are susceptible to erosion due to the flow of water. Conventional approaches to scour analysis, encompassing physical model tests and empirical equations, tend to be time-intensive, costly, and prone to inaccuracies. As machine learning (ML) algorithms have progressed, novel avenues for conducting scour analysis have emerged, presenting more efficient and precise solutions. This review article aspires to furnish a comprehensive overview of recent advancements in scour analysis through the utilisation of ML techniques. The article's outset involves a definition of scour analysis and an explication of established methodologies in this realm. Subsequently, the discussion pivots to outlining the merits of deploying ML techniques for scour analysis. The article delves into a survey of diverse ML methods employed for scour analysis, encompassing supervised learning as well as unsupervised learning algorithms. Furthermore, the article includes instances of ML techniques employed in scrutinising scour analysis concerning bridge piers. The review culminates in a discourse on challenges encountered and prospects on the horizon within the domain of scour analysis via ML techniques. The challenges identified encompass the dearth of high-quality data for training ML models and the intricacies associated with comprehending ML models. Anticipated progressions in ML techniques for scour analysis could be geared towards enhancing the intelligibility of models and tackling the hurdle of limited data availability. In a holistic sense, this review article spotlights the latent potential of ML techniques in refining the scour analysis process, all while pinpointing avenues for prospective exploration.