Abstract: In order to improve the numerical stability and computation efficiency and prevent the divergence of adaptive Ksman fillter, a new robust adaptive extended Kalman filter based on U-D covariance fsctorization is proposed in this paper. The new algorithm is used in the flight state estimation of aircraft. The results of simulated and real flight test data computation show that the new algorithm proposed can provide more precise estimation than the extended Kalman filter in solring flight state estimation problem corrupted by timevariant noise. Moreover, the new algorithm requires less noise statistics information, and has better convergency than the adaptive extended Kalman filter. When the adaptive extended Kalman filter diverges, the new algorithm can also give good estimation results.