The use of the vehicle information system and the complex road traffic control information make the mental workload of drivers increased. In order to recognize driving mental workload efficiently, provide the basis of automatic auxiliary driving and integrate the traffic information, the method use the EEG signal δ(0.5-4 Hz), θ(4-8 Hz), α(8-13 Hz), β (13-30 Hz) as the input features and SVM model to establish the recognition model for state of driving mental workload. Meanwhile, combine with examples based on EEG data from the simulator to test the model, the result shows that the average recognition accuracy rate was between 93. 8% and 96.5%. The modle shows good accurancies for driver's mental workload recognition and can be used in actual driving.