Objective To better meet the accuracy requirement of vehicle heading angle measure.Method Kalman filter algorithm was proposed. The latitude, longitude and elevation measured by RTK-GNSS receiver were converted to plane coordinates using Gaussian projection. The Gaussian plane coordinates and the accumulated heading angle measured by gyroscope were integrated by Kalman filter, and finally more accurate heading angle was obtained.Result The integrated curve kept the entire variation trend of the heading angle measured by GNSS receiver and the partial variation trend measured by MEMS gyroscope. The curve was smoother than that based on GNSS and gyroscope, and could follow action of vehicle 180° turning.Conclusion The Kalman filter algorithm can measure the vehicle heading angle data in real time and the precision was improved 80% more than the result measured by GNSS.