This article reviews mainly the advancement of artificial intelligence technology in stroke screening, and analyses the research directions of intelligent assessment algorithms based on the establishment of stroke screening by using neural networks and deep learning algorithms. Artificial intelligence application has been emerging in both regular population screening of stroke and screening and prevention of stroke for a high-risk population. At present, the main problems, however, are the small amount of accumulated screening data, the lack of standardized and AI-suitable datasets, and the discrepancy in predictive accuracy between AI-assisted screening and traditional analysis methods due to particular data and undefined risk factors. Therefore, we should establish standardized datasets and adopt deep learning algorithms suitable for structured feature data for intelligent screening research of stroke.