This paper provides a comprehensive survey of domain generalization (DG), a field that focuses on learning models that can generalize well to unseen test domains. The authors define DG and discuss related fields such as domain adaptation, transfer learning, and meta-learning. They categorize recent algorithms into three classes: data manipulation, representation learning, and learning strategies. The paper also introduces commonly used datasets, applications, and an open-sourced codebase for fair evaluation. Finally, it summarizes existing literature and presents potential future research directions. The survey covers theoretical foundations, methodological advancements, and practical applications, making it a valuable resource for researchers and practitioners in the field of machine learning.This paper provides a comprehensive survey of domain generalization (DG), a field that focuses on learning models that can generalize well to unseen test domains. The authors define DG and discuss related fields such as domain adaptation, transfer learning, and meta-learning. They categorize recent algorithms into three classes: data manipulation, representation learning, and learning strategies. The paper also introduces commonly used datasets, applications, and an open-sourced codebase for fair evaluation. Finally, it summarizes existing literature and presents potential future research directions. The survey covers theoretical foundations, methodological advancements, and practical applications, making it a valuable resource for researchers and practitioners in the field of machine learning.