This paper is concerned with the estimationof heading information of the Pedestrian Navigation System (PNS).The MEMS inertial sensors as well as a miniature GNSS receiver are used to establish a pedestrian navigation prototype based on the Pedestrian Dead Reckoning (PDR) approach.An Extended Kalman Filter (EKF) structure is used for the estimation of the system’s attitude error and the bias of the gyroscope.If there is no external acceleration,errors of pitch and rollas well as the biases of the two horizontal gyros are compensated using the aiding information from the accelerometer.When GNSS is available,its output is usedfor heading and heading-gyro bias estimation.Using the aiding information provided by both GNSS and accelerometer,the proposed method prevents the errors in the attitude from increasing rapidly.The proposed methodfor pedestrian navigation application has been well verified through real field experiments.