To solve the problems of low yaw observability low attitude estimation precision and easily diverged estimated angles in the integrated navigation a low-cost attitude estimation algorithm based on the integrated navigation aided by the yaw of the single baseline GPS is proposed.The model adopts integrated navigation estimation model with 9 states as the system states.A method of conversion between the quaternion and the Rodrigues parameters is adopted to avoid the weighted average calculation of the quaternion and the single covariance.A method of assisted GPS track angle is proposed to improve the observability of yaw angle and the performance of the algorithm.An adaptive UKF is proposed to improve the applicability of the algorithm and overcome the low precision and easily diverging problems of the EKF.As a result accurate and real-time attitude estimation is eventually realized.The precision of the algorithm is higher than that of the EKF.The maximum estimation error is within 0.5°, and the RMS of attitude estimation values are less than 0.36° which meets the requirement of the beam alignment for Mobile Satellite Communications.