The problem of identifying the modulation types of the signals at the receiver is an intermediate step between signal detection and demodulation. In this paper, Hilbert-Huang Transform is proposed for identifying the modulation level of M-ary Quadrature Amplitude Modulation (QAM) signals in the presence of the Additive White Gaussian Noise. Hilbert-Huang Transform decomposes the non-stationary signal into sum of oscillatory signals with different frequency and obtains the instantaneous features of the signals, like time-frequency and time-amplitude information. The statistical properties of instantaneous data sets are used to determine the modulation levels of M-ary QAM signals. The most important feature of the proposed method is that, the algorithm does not utilize any a priori knowledge about the signal. Computer simulations demonstrate the accuracy of the proposed algorithm.