Surfactants, as a class of important functional compounds, are widely used in daily chemicals, biomedicine, petroleum extraction, and other fields. However, the traditional trial-and-error experimental approach not only struggles to meet the performance requirements of novel surfactants under extreme conditions but also lacks systematic evaluation of their environmental fate and biosafety. Machine learning, as a data-driven research paradigm, offers a novel approach for efficient surfactant development, demonstrating significant potential particularly in molecular property prediction. This article systematically reviews advances in machine learning for predicting key surfactant properties (e.g., critical micelle concentration, surface tension, hydrophilic-lipophilic balance, adsorption efficiency, and Krafft point). Furthermore, it examines preliminary applications and prospects of this technology in predicting environmental safety parameters such as biodegradability and ecotoxicity. Current research predominantly focuses on predicting critical micelle concentration, surface tension, and hydrophilic-lipophilic balance, whereas modeling studies on adsorption efficiency, Krafft point, and environmental safety parameters remain relatively limited. By extending predictive capabilities to these underexplored properties, machine learning is poised to become a pivotal tool for rational design, performance optimization, and green assessment of surfactants, thereby advancing the development of high-performance and environmentally friendly surfactants.

Key words: Machine learning, Surfactant, Property Prediction, Molecular Representation, QSPR