Abstract: After reviewing the application of system identification to aircraft flutter research, three methods of system identification are presented in this paper. Emphases are put on the main points of identification algorithm and its theoretical basis. The problems related to data processing are also discussed. The Transfer Function Method based on complex modal analysis and optimization technique can identify all of the modal parameters accurately. Furthermore, this method has been applied to the flutter model test of a wing with external bodies successfully, The method can be extended to the cases of wind tunnel and flight tests which provide response data only. Adopting the correlation and least square technique, Impulse Function Method and Autoregressive Moving Average Method can gain considerable accuracy in identification of measurement data conteminated with rather strong noise interference. In no need for a special signal analysis instrument, it is possible to make direct use of measurement sampled data from artificial random or transient excitation and response data from natural random excitation with these two time domain methods.