In order to solve time-variable of a non-steady random signal during the process of structure health monitoring and identify the damage efficiently,a progressive damage identification method for the single degree of freedom (DOF)structure is studied based on Hilbert-Huang transform(HHT).Firstly,vibration acceleration signals of a single-DOF structure model were generated by reducing the stiffness gradually.Secondly,the signals were decomposed into several smooth intrinsic mode function (IMF),some IMFs which contain dominant damage information were selected and transformed by HHT.Finally,the instantaneous frequency was calculated as the feature parameter.The result indicates that the instantaneous frequency is sensitive to damages and changes obviously after the damage,which can be taken as a feature index to monitor the progressive structural damage.