The application of conventional Kalman filtering to initial alignment of strap down inertial navigation system(SINS) may influence the estimation effect due to the uncertainty of statistic characteristics of the model parameters and noise. The fuzzy adaptive Kalman filtering algorithm can modify the measurement noise covariance matrix gradually through Fuzzy Inference System (FIS). The detailed method is that the weighting of the observing noise covariance is adjusted by the system to correct the measurement noise covariance by observing whether or not the theoretic value of the residual closes to the real measurement covariance, and then the alignment efficiency of the navigation system can be improved. When the noise statistic characteristics are unknown,the initial alignment effects of the conventional Kalman filtering and the fuzzy adaptive Kalman filtering have been compared.The simulation results show that the proposed algorithm can improve the filtering performance of the navigation system effectively,and it is an ideal navigation filtering method of initial alignment.