Abstract: Single-channel seismics, as a common method, is widely applied in marine geophysical survey. However, the method has some disadvantages, such as low signal-noise ratio, interference of multiple wave and large difference in vertical amplitude caused by single coverage and unequal acquisition intervals. Based on the acquisition principle of single-channel seismic survey, this paper is devoted to the three difficulties in processing single channel seismic data, i.e. seafloor in-phase axis jitter, seafloor multiple waves with strong amplitude, and serious background random noise. In order to solve the problem of seafloor in-phase axis jitter caused by surge wave, we subdivided the seafloor jitters into high frequency and low frequency components according to the principles we usually used for land data static correction. Cross-correlation technique is used to gradually eliminate the high and low frequency static correction components. For multiples, one-dimensional SRME algorithm formula is applied to single-channel data processing and make the predicted multiple wave model close to the actual. Concerning the problem of low SNR of single channel data, RNA technology combined with LIFT theory is adopted to enhance signals and suppress spatial random interferences. The SNR and resolution of the actual data after processing are significantly improved, and the wave group features become clearer.