A nondestructive evaluation technique established on the basis of a global minimization method is presented for the system parameters identification of flexibly supported rectangular laminated composite sandwich plates using measured natural frequencies of the sandwich plates. In this study, the first eight natural frequencies extracted from impulsive vibration testing data are used to identify the system parameters of the flexibly supported sandwich plates. In the identification process, the trial system parameters are used in the Rayleigh-Ritz method to predict the theoretical natural frequencies of the sandwich plate, a frequency discrepancy function is established to measure the sum of the squared differences between the experimental and theoretical predictions of the natural frequencies, and the global minimization method is used to search for the best estimates of the system parameters by making the frequency discrepancy function a global minimum. The accuracy and efficiency of the proposed technique in identifying the system parameters of several flexibly supported sandwich plates made of different face and core materials are studied via both theoretical and experimental approaches. The reasonably good results obtained in this study have demonstrated the applicability of the proposed technique.