Antifouling surfaces, renowned for their strong surface resistance to proteins, cells, or tissues in various biological and environmental conditions, have broad applications in implanted devices, antibacterial coatings, biosensors, responsive materials, water treatment, and lab-on-a-chip. While extensive experimental research exists on antifouling surfaces, machine learning studies on this topic are relatively few. This perspective specifically focuses on exploring the complex relationships between the composition, structure, and properties of antifouling surfaces, examining how these factors correlate with surface hydration and protein adsorption. Different machine learning models have been developed to analyze and predict single and multiple protein adsorptions on various types of surfaces, ranging from structureless surfaces to well-ordered and rigid self-assembled monolayers, dynamically ordered polymer brushes, and complex filtration membranes. These models not only identify key descriptors or functional groups critical for antifouling performance (surface hydration, protein adsorption) but also predict the antifouling properties for a specific surface. Recognizing current challenges, this perspective delineates future research directions in the antifouling field. By leveraging and comparing current machine learning approaches, it aims to advance both the design and fundamental understanding of antifouling surfaces, thereby pushing the boundaries of innovation in this critical field.