In this paper, a system for automatic recognition of radar waveform is introduced. This technique is used in many spectrum management, surveillance and cognitive radio and radar applications. For instance the transmitted radar signal is coded into six codes based on pulse compression waveform such as Linear Frequency Modulation (LFM), Frank code, P1, P2, P3 and P4 codes, the latter four are poly phase codes. The classification system is based on drawing Choi Willliams Distribution (CWD) picture and extracting features from it. In this paper, various new types of features are extracted from CWD picture and then a pattern recognition method is used to recognize the spectrum. In fact, signals from CWD picture are defined using biometric techniques.