The invention discloses a ship behavior analysis method based on a YOLO model. The method comprises the steps of manufacturing a data set of the existing ship type images; extracting the multi-scale features of the ships in a data set through a convolutional neural network of the YOLO model, and training the ability of the neural network to identify the ship images; taking a video sequence of theships as an input parameter of the YOLO model, identifying the ships in the video sequence, and obtaining the ship position information; and according to the position of the ship to be detected in each frame of the video sequence, analyzing the behavior of the ship to be detected by applying the space-time constraint conditions and the geometric knowledge. The method provided by the invention hasthe higher robustness, can analyze the ship behaviors, can identify the historical ship behaviors, and is helpful for the maritime work personnel to identify and predict the potential dangerous ship behaviors, so that the early active activities are adopted to improve the maritime traffic safety.