The installation bias angle error and the scale factor error of the acoustic Doppler log are the main error sources that affect the underwater autonomous navigation accuracy，which needs accurate error calibration. However，any sonar performance is closely related to the marine environment. How to achieve accurate calibration in complex environment is a practical problem for log engineering application. Therefore，based on the analysis of the statistical distribution of real velocity measurement data，this paper proposes an error calibration method based on modified Kalman filter preprocessing，and compares it with the standard Kalman filter preprocessing method in terms of calibration performance. The simulation and test results have verified the robustness of the proposed method， effectively reduced the requirements for calibration test site，and greatly improved the calibration efficiency.
