Abstract: It is well known that adaptive line enhancer (ALE) can be used to effectively detect narrow-band signal or CW pulse signal with unknown frenqency in presence of noise However, it has been shown that the processing gain of adaptive line enhancer using LMS algorithm is not satisfied because of the presence of iterative noise and weight noise. In this paper, a coherent accumulative algorithm of adaptive line enhancer is proposed. One result of this analysis is that adaptive LMS filter possesses the ARMA model construction. So more powerful ability to filter out noise can be obtained. The experimental results carried out in laboratory show that the processing gain of this new arthitectureal ALE is about 14 dB higher than that of conventional ALE.