Abstract：To correctly classify advanced radar emitter signals, a novel approach using image feature of Choi-Williams time-frequency distribution for radar emitter signal recognition is proposed, which transforms the classification of emitter signals into image processing and image recognition. Time-frequency images of radar Emitter signals are obtained by using Choi-Williams distribution, and then these images are transformed into grayscale images. A series of image processing methods are employed for time-frequency image enhancement and de-noising, and the grayscale images are converted into binary images following. In addition, the areas not containing signal components from the edges of the image are removed. Finally, the centralize moments and pseudo-zernike moments are calculated as the feature for signal recognition, and the support vector machine is used to identify radar emitter signals automatically. Simulation results show that the proposed approach can achieve satisfying accurate recognition when signal-to-noise rate (SNR) varies in a large range. Even for SNR=-3dB, the proposed method which adopts pseudo-zernike moments works effectively as high as 92% recognition rate . The validity of the approach is demonstrated by experiments.