Aiming at the problem of blind estimation of pseudo-noise codes in multi-rate long and short codes direct sequence code division multiple access signal, in this paper, a novel codes estimation method based on Fast-ICA algorithm and sequences properties is proposed in this paper. The received signal is firstly segmented twice according to its maximum long scrambling code period and minimum spreading code period. Each user’s composite code fragments consisting of long and short codes are separated by Fast-ICA algorithm. Then the user's long and short codes are estimated in descending order of the data rate, and the specific steps are as follows. Firstly, the separated composite code fragments make up the fuzzy sequences, and the double delay-and-multiply method is used to eliminate the order fuzzy and spread code interference. Secondly, combined with cyclotomic cosets and triple correlation properties, the method of feature information matching is used to estimate the long scrambling codes of all users with the same data rate. Meanwhile, the short spread codes are estimated by correlation operation. Lastly, all the estimated users’ composite code fragments are deleted by using similarity matrix, fuzzy sequences are reconstructed, and three steps above are repeated until all users’ long and short codes are estimated. Simulation results show the effectiveness of the proposed method. © 2019, North Atlantic University Union. All rights reserved.