Recognition of low probability of interception (LPI) radar signals is extremely important in electronic support measurement (ESM), or electronics intelligence (ELINT) systems. Polyphase codes including Frank, P1, P2, P3, and P4 code are a kind of LPI signals, which have been a new research hotspot in reconnaissance systems. In this paper, we propose a new recognition method for polyphase coded signals using the time-frequency rate (TFR) distribution. First we use the property of linear frequency modulation (LFM) to distinguish polyphase code from other LPI signals. Then we analyze the differences between polyphase code and LFM, and study the recognition feature of peak slice of the TFR. Finally we propose a hierarchical scheme for the recognition of polyphase coded signals. The simulation experiments results demonstrate that the proposed method has a high accurate recognition rate for various polyphase code, even under low signal to noise ratio (SNR).