Abstract: A discrimination method between submerged and surface targets is proposed based on wavenumber energy frequency domain extension features. The frequency wavenumber spectrum is calculated through modal domain beamforming, and a mapping relationship between target depth and wavenumber extension differences is established based on environmental information, which is used as a decision model to achieve target depth identification. The simulation results show that the proposed method is still robust when the target azimuth estimation is not accurate. The effectiveness of the method was verified through SWellEx-96 data and measured data in the Dalian sea area. The processing results of sea trial data showed that the method has preliminary ability to distinguish surface and submerged targets under the conditions of a 132 m horizontal array longitudinal effective aperture and 100−800 Hz processing frequency band. The wavenumber expansion feature of the surface target is about 10° higher than that of the submerged target, and the difference between the two types of targets is significant. From the perspective of sound propagation characteristics, the proposed approach provides new solution to the problem of target depth discrimination.