Empirical mode decomposition (EMD) and wavelet decomposition (WD) are two new time-frequency analytic methods used to analyze the non-stationary signal. Each of these two analytic methods holds its characteristics in non-stationary signal processing. A contrast experiment among simulated experiments of signal mutation, signal trend extraction and frequency detection demonstrated that WD was better than EMD in signal mutation detection, while EMD was better than WD in low frequency signal detection and signal trend extraction.oft-com:office:office" />