The invention relates to a vehicle / vehicle flow detection method based on a three-axis magnetoresistive sensor, comprising the following steps: calibrating or calibrating the three-axis magnetoresistance sensor; obtaining the environmental magnetic field strength when there is no car, and obtaining the magnetic field strength when there is a car passing by; using Kaiser The FIR filter is designed to filter the measured magnetic field strength data to obtain the data sequence; the data sequence is binarized; whether there is a car is judged; when there is no car, interference is judged; when there is a car, interference is judged; the vehicle leaves the judgment; after the vehicle is judged The software forcibly resets the three-axis magnetoresistive sensor. It is characterized in that the method integrates the three-axis magnetic field information, which can fully reflect the disturbance information of the geomagnetic field when the vehicle passes by, the sensor can be placed arbitrarily, and the software reset method can effectively overcome the signal drift of the magnetic resistance sensor generated when the vehicle passes continuously Problems, through the judgment of vehicle entry, interference, and vehicle departure, it can ensure high detection accuracy and strong anti-interference ability. This method is easy to implement and is suitable for vehicle / traffic flow detection on traffic roads, bridges, squares, tunnels, etc. It can also be used in Vehicle detection in the parking lot for intelligent management.