Modulation detection, which is one of the major tasks of an intelligent receiver in a cognitive radio communication systems, is an intermediate step between signal detection and demodulation. In this paper, we present a time-frequency based method for the detection of modulation type in the presence of noise. The Discrete Evolutionary Transform is used to distinguish PSK and FSK signals. The performance of the proposed method is compared with one of the Wavelet Transform based methods in the literature. Simulation results confirmed the virtue of the proposed method.