<p>The availability of extensive literature on quantifying the mechanical behavior of engineered structures made of functionally graded materials (FGMs) highlights their importance. Traditional methods, such as analytical solutions, discretized approaches (e.g., FEM), meshless methods, semi-analytical methods, and hybrid approaches, have been widely employed under equivalent single-layer, layerwise, or elasticity theories to study FGMs under diverse loading conditions. However, with the rise of artificial intelligence (AI) and its ability to process large datasets efficiently, machine learning (ML) has emerged as a transformative tool for analyzing FGM behavior. This review systematically consolidates existing research on the Machine Learning-Driven prediction of mechanical properties and structural responses in functionally graded materials (FGMs). It provides a detailed taxonomy of algorithms (e.g., neural networks, surrogate models), data sources (experimental, simulation, or hybrid datasets), input variables (e.g., gradation profiles, loading conditions), and output metrics (e.g., stress distribution, natural frequencies). Challenges associated with ML surrogates—such as data scarcity, interpretability, and computational constraints—are critically reviewed. This review aims to link advancements in ML and FGM mechanics, serving as a benchmark to identify gaps in existing knowledge and catalyse future research in this novel intersection where ML draws from the mechanics of FGM, providing a pathway to next-generation predictive modeling in the realm of smart material design.</p> Graphical Abstract <p></p>