A multi-input multi-output (MIMO) method for identifying flight vehicle structure modal parameters was proposed based on the continuous wavelet transforms. The auto regression(AR)model in the wavelet domain was established by applying the continuous wavelet transform to discrete equations of motion. The coefficient matrices of the AR model, which determine the dynamic characteristics of the system, were calculated through the least squares approach. The modal parameters were then computed by solving eigenvalues of eigenmatrix constructed by these coefficient matrices. Optimization algorithm was used to improving accuracy and stability of proposed approach in process of identification. The advantages of the proposed approach over the exiting methods of applying the wavelet transform to modal parameters identification are in use of the time invariance property and filtering ability of the transform to enhance the efficiency of identification. Simulation result shows this method is accurate and stable and can be used to identify flight vehicle structure modal parameters.