Abstract: In order to solve the problem of spread spectrum signal distortion caused by raised correlation between the samples with general time-domain adaptive filtering narrowband interference suppression method in the oversampling direct sequence spread spectrum (DSSS) system, an improved time-domain adaptive filtering algorithm was proposed based on interpolating transversal filter with two-sided taps and the minimum mean square (LMS) algorithm to update the weights adaptively. Combined with downsample-upsample structure, it could effectively suppress narrowband interference. Both theoretical analysis and simulations show that, compared with traditional adaptive filter method, the proposed algorithm can not only lower signal-to-noise ratio (SNR) loss, but also get excellent characteristics as small distortion of desired signal in high SNR conditions, meanwhile it has a slight advantage in bit error rate(BER) performance in the low SNR conditions.