The laser strapdown inertial measurement unit(LSIMU) is greatly interfered by measurement noise in the working environment, especially the small laser gyroscope, which cannot accurately identify the angular velocity of the earth''s rotation. As a result, the strapdown inertial navigation system cannot accurately and quickly perform initial alignment on the moving base. To solve this problem, an initial alignment algorithm of LSIMU moving base based on wavelet Kalman cascade filter is proposed in this paper. The results of simulation and comparison experiments show that compared with other initial alignment algorithms, this algorithm can effectively improve the initial alignment accuracy, reduce the attitude angle alignment error. The initial alignment accuracy of the global positioning system assisted laser strapdown inertial navigation system increases from 10'' to 5'' with an minimum variance, which can make the alignment error more stable and fast convergence.