Motion Style Transfer (MST) is a key research area in computer graphics, enabling motion sequences to acquire stylistic attributes while preserving semantic consistency. Recent advancements have significantly improved MST’s flexibility and scalability, facilitating high-quality motion adaptation across various applications. This paper presents a comprehensive survey of MST research, categorizing existing approaches into generative models, content-target combination methods, and text-semantics-driven techniques. We further review commonly used MST datasets and evaluation metrics, highlighting inconsistencies in benchmarking methodologies. We highlight the need for scalable dataset generation and standardized evaluation protocols to advance MST research.