Abstract: A blind estimation method for pseudo noise (PN) sequence estimation of aperiodic direct sequence spread spectrum (DS) signals under condition of lower signal noise ratio (SNR) is presented. The proposed method estimates the PN sequence with segmented matrix eigenvalue decomposition (SEVD) method whose segment length equals double information bits,and then shells off the random information bits from the estimated sequences to obtain the original PN sequence based on the optimal shift-and-add property. The method accelerates the computation speed by applying parallel computing without need of prior synchronization information,and it can sweep off the “code blurring” phenomena. Simulation results show that the aperiodic DS signal with m sequence period 1 023 can be estimated without any errors as long as the SNR is above ?17 dB.