Abstract: Marine vertical cable seismic (VCS) is an unconventional exploration method that acquires seismic data through vertically deployed hydrophone arrays, enabling high-resolution imaging of subsea geological structures. However, due to the limited number of acquisition channels in practical operations, conventional wavefield separation methods—such as f-k filtering and radon transform—often suffer from insufficient spatial sampling, severe dispersion, and waveform distortion, making it difficult to achieve high-precision separation of upgoing and downgoingwavefields.To address this challenge, this paper introduces a differential-equation-based filtering (DEBF) method for wavefield separation in VCS data processing. The approach involves designing a zero-phase filter in the frequency-wavenumber (f-k) domain, transforming it into the time-space domain, and implementing stable and efficient wavefield separation using finite-difference algorithms. This method demonstrates excellent amplitude fidelity and adaptability. Tests on both synthetic and real datasets show that the proposed method effectively separates upgoing and downgoing wavefields even with limited channel counts, significantly outperforming conventional techniques in terms of separation quality and reliability.