The invention relates to an SVM and Bayesian network FlexRay bus fusion intrusion detection method and detection device. The detection method comprises the following steps: collecting a sample to be detected; carrying out sample pretreatment on the collected to-be-detected sample; importing the preprocessed sample into a trained SVM detection model for detection; judging a result detected by the trained SVM detection model, and if it is judged that intrusion does not occur, ending intrusion detection; if it is judged that intrusion occurs, a judgment result is imported into the trained Bayesian network detection model, and an intrusion source is distinguished; the system sends an intrusion alarm and reports a source, and intrusion detection is ended; the vehicle-mounted FlexRay bus safetyprotection system is used for providing safety protection guarantee for the vehicle-mounted FlexRay bus; the intrusion behavior in the bus data is detected, an intrusion source (a physical layer or anetwork layer) is judged, a network administrator is informed of a detection result to carry out corresponding processing, the detection result can prevent attacks in advance, and technical support isprovided for information safety protection of the intelligent networked vehicle.