Abstract: An improved synchronization iterative ML algorithm based on cyclic prefix was presented. This algorithm considered the effect of multi-path fading channel and took the advantage of the sample points with the same position in many OFDM symbols added each other and amended these sample points to compute their correlation. Therefore it can reduce the variety of amplitude of the sample points and improve precision of the proposed synchronization iterative algorithm. It can also improve the weakness of the primary iterative ML algorithm in the estimation of timing and frequency offset in AWGN noise channel and establish accurate synchronization in multi-path fading channel. Simulation results conducted in both AWGN channel and multipath fading channel show that the improved algorithm is characterized by better synchronization performance.

Key words: information processing, OFDM systems, iterative ML algorithm, timing and frequency offset, synchronization, cyclic prefix