Abstract: To decrease the BER (bit error ratio) of MMSE (minimum mean square error) channel estimation algorithm for OFDM systems, an improved channel estimation algorithm by using band-stop filter is proposed. In the proposed algorithm, the initial estimate value of channel is given by MMSE. Then, the noise is separated from channel frequency response by use of the characteristics of DFT (discrete Fourier transform). Finally, the channel estimation algorithm is filtered by the band-stop filter so as to improve the system accuracy of BER performance. Simulation results showed that the performance of the proposed channel estimation algorithm is obviously better than classical MMSE channel estimation algorithm on condition that the pilot frequency interval remains unchanged or increases gradually, though the computational complexity is increased slightly.