Abstract: The time domain signal of the SRS(Shock response spectrum) is a typical non-stationary signal, and the orthogonal HHT (Hilbert-Huang Transform) method is used for the time-frequency analysis of non-stationary signals. The time domain signal measured from the shock response spectrum test machine is selected as a case, the time-frequency characteristics are analyzed by the orthogonal HHT method. Then combined with the advantages of the evolutionary spectra theory and the orthogonal HHT method, a method for the simulation of the shock response signal is proposed. It is shown that for the synthetic sample, the time-frequency distribution is consistent with the original shock response signal, and is better than the traditional triangle wave synthesis method in non-stationary process simulations. At last, based on a large number of synthetic samples, a mathematical model of the SRS and its corresponding fitting parameters are presented; and the results can provide a useful reference for the related fields.