Aiming at the problem that the domestic civil aircraft navigation data can not be accurately measured due to clutter, a navigation data filtering method based on improved adaptive unscented Kalman filter (AUKF) algorithm is proposed. The improved AUKF is constructed by combining unscented Kalman filter (UKF) with improved sage-Husa sub optimal unbiased maximum posterior noise estimator. The problem of low accuracy and divergence is effectively solved in the case of uncertain model or incomplete knowledge of statistical characteristics of interference signals. At the same time, the filter effect is compared with Wiener filter and wavelet threshold method. The simulation is carried out with the longitude, latitude and altitude data of domestic ARJ21. The results show that the filtering accuracy of the improved AUKF algorithm is higher than others, and effectively improves the reliability of navigation data. The research is of great significance to improve the navigation and positioning accuracy of domestic civil aircraft.