The invention relates to an autonomous underwater vehicle (AUV) push-position navigation system bar arm error calibration method, and belongs to the field of an underwater vehicle. According to the AUV push-position navigation system bar arm error calibration method, rod arm errors due to the fact that in a push-position navigation system, a doppler velocity log (DVL) is not installed on the center of mass of the AUV together with an attitude and heading reference system (AHRS) are calibrated, an error calibration model suitable for a kalman filtering algorithm is built, and an easy-to-operateturning movement way of identifying a rod arm length is presented. Heaving information is not needed to be provided by the AHRS, the requirements for inertial base navigation equipment are reduced, ascheme has a good universality, and is suitable for solving rod arm length identification problems of various types of AUVs, and the positioning accuracy of the AUV push-position navigation system can be effectively improved by compensating the rod arm errors in the AUV push-position navigation system; and meanwhile, the scheme can be directly obtained, the situation that the DVL, the AHRS and the center of mass (or center of buoyancy) are located on the same straight line is not required, that is to say, no matter where the center of mass (the center of buoyancy) of the AUV is located, the corresponding rod arm length can be obtained directly, and the application prospect is wide.