Continuous wavelet transformation is made to identify the parameters of damped harmonic forced vibration Duffing system. With the aid of conversion relationship between the scale and frequency, the solution of nonlinear Duffing equation is adopted by average method, which gained approximate analytical expression for instantaneous amplitude and instantaneous frequency of the system. The nonlinear stiffness coefficient and natural frequency can be gained by least square method and the relationship between recognition accuracy and parameter selection are summarized in the article. Parameter identification method of harmonic forced vibration system is proposed in this paper. Studying the wavelet ridge and corresponding scale by segments to filter out the affects of the simple harmonic motion, to extract systems free vibration signal and to achieve the goal of identifying system parameters.