Abstract: An underwater acoustic imaging method for pattern recognition in a free environment was proposed. Angular domain filtering was used to separate plane waves from different direction of space in the array domain. According to requirements in the imaging sector, a series of angular domain band-pass filters were designed, and outputs of these filters were the snapshot data from an incident plane wave within the angular pass-band in the array element domain.Some special pattern recognition algorithms, such as target recognition and bottom classification, which needed the characteristic data of the array domain, could be realized. The amplitude of snapshot data from output of angular domain filter as the pixel value could be realized underwater acoustic imaging. The method of angular domain filtering could be used for underwater acoustic imaging so obtained the desired sound image, and each pixel corresponded to a snapshot data generated by the corresponding target far-field plane wave which was incident on the receiving array.The results of simulation and experimental data processing showed that the sound image obtained by angular domain filtering was same as that used the conventional beam forming method, and two imaging methods had same robustness.