Abstract: Corrosion inhibitors are chemical substances used to prevent the corrosion of metallic materials, and their effectiveness is crucial for extending equipment lifespan and reducing maintenance costs. However, traditional screening methods for corrosion inhibitor molecules, such as weight loss measurements and electrochemical testing, typically require extensive experimentation and time, leading to high costs. In recent years, machine learning (ML) has demonstrated significant potential in the field of corrosion inhibitor research, emerging as a powerful tool for scientists exploring new, efficient inhibitors. This review first provides a brief overview of the application of ML techniques in the screening of corrosion inhibitor molecules. Specifically, ML models can be employed to select descriptors of corrosion inhibitor molecules through data preprocessing and feature extraction, thereby constructing predictive models for inhibition efficiency (IE). We categorize the IE predictive models into those based on molecular descriptors and those based on molecular structures, both capable of rapidly identifying potential high-efficiency corrosion inhibitor molecules, significantly accelerating the research process. However, despite the advantages of ML in the screening of corrosion inhibitor molecules, certain limitations must not be overlooked. Firstly, the current screening models have a limited chemical search space; secondly, practical applications of these models face challenges related to computational resources and time costs. Consequently, this paper further explores the concept of molecular generation techniques and their prospects and challenges in the task of generating corrosion inhibitor molecules. Overall, through continuous optimization and improvement of ML algorithms, combined with experimental validation, the future holds promise for the discovery of efficient corrosion inhibitors, potentially bringing new breakthroughs to materials science and industrial applications.