In view of the difficulties of vehicle trajectory prediction, the classification and research status of vehicle trajectory prediction methods are reviewed. According to the different prediction time domain of model implementation, the existing algorithms are divided into short time domain and long time domain vehicle trajectory prediction methods; the basic concepts and research status of short time domain prediction methods based on physical model and traditional machine learning are introduced, and the long time domain prediction methods based on deep learning, neural network and vehicle driving behavior intention recognition are summarized and compared. The analysis results show that the long time domain method can solve the difficult problem of vehicle trajectory prediction and ensure the efficient and safe driving of intelligent vehicles.