Abstract: In order to solve the problem of the background noise in the communication signals, the preprocessing method based on digital morphology can filter the background noise well. Because it is difficult to detect the interferences of different widths with a signal threshold, a narrow-band interference detection algorithm based on morphological adaptive threshold is proposed. Firstly this algorithm estimates the spectrum of signals, and then preprocesses them by morphological method. Then, according to the distribution of signal power spectrum, different threshold values were selected to realize adaptive threshold, which provided an effective method for detecting narrow-band interference of different signal bandwidths. The method proposed in this paper is not affected by the change of the noise base, and it has low computation and complexity, which is suitable for the real-time spectrum monitoring of on-board satellite communication. Through Matlab simulation experiments, it is concluded that the detection effect of flat structural elements with a length of 25 is improved by more than 6dB compared with the traditional continuous mean removal algorithm (CME) through the expansion preprocessing method and adaptive threshold.