Abstract: Aiming at the problems of poor indoor positioning accuracy and complex implementation, a pedestrian indoor positioning algorithm based on pedestrian trajectory estimation (PDR) is proposed. The algorithm calculates user location by the raw data which is gathered by the phone built-in accelerometer of inertial sensor, gyroscope and magnetometer. It gets stride frequency and step size of pedestrian by utilizing calculation of acceleration data and acquires attitude and heading by using quaternions to solve gyroscope data.Then, it calculates the more accurate course information by utilizing extended kalman filter to blend all sensor data and finally gets pedestrian location by renewal of location. By sets of experiment, the precision of indoor position is superior to 1.8 percent. The algorithm has no need of arranging additionally beacon nodes and has higher position precision, lower complexity of realization and higher usefulness.