The geological conditions for coal mining in China are complex, with various structural issues such as faults and collapsed columns seriously compromising the safety of coal mine production. In-seam wave exploration is an effective technique for acquiring detailed information on geological structures in coal seam working faces. However, the existing reflected in-seam wave imaging technique can no longer meet the exploration precision requirements, making it imperative to develop a new reflected in-seam wave imaging technique. This study applies the Gaussian beam summation (GBS) migration method to imaging coal seams' reflected in-seam wave data. Firstly, with regard to the characteristics of the reflected in-seam wave data, methods such as wavefield removal and enveloped superposition are employed for the corresponding wavefield separation, wave train compression and other processing of reflected in-seam waves. Thereafter, imaging is performed using the GBS migration technique. The feasibility and effectiveness of the proposed method for reflected in-seam wave imaging are validated by conducting GBS migration tests on 3D coal-seam fault models with different dip angles and throws. By applying the method to reflected in-seam wave data for an actual coal seam working face, accurate imaging of a fault structure is obtained, thereby validating its practicality.