Abstract: The detection of slow-moving target is seriously hampered by ionospheric contamination in the over-the-horizon radar(OTHR). The clutter spectrum is broadened worse when the phase disturbance and multi-mode propagation coexist. In this paper, a method for suppressing multi-mode clutter based on eigenvalue decomposition is presented to solve the problem mentioned above. The echo signals of some group-distance cells selected in according to the correlation among different cells are used to construct the covariance matrix, and the mode number of echo signal is obtained through sliding filtering using a template filter. After that, the target signal and noise subspace is determined, and then, the multi-mode clutter is suppressed. The method is easy to implement, thus suitable for engineering applications. The processing resuleness of the measured data show the effectiveness of the proposed method.