Abstarct: Nowadays, it is very important to obtain dynamic specifications to monitor the health and control and improve the structures that were built in the past. Over the years, a variety of methods baxsed on signal processing knowledge have been proposed that have recently made significant progress.these methods has limitations that some of blind source separation limitation have been addressed in this dissertation. In structures with high damping, identification of mechanical properties such as natural frequency, damping ratio and modal shapes is very important because the presence of energy depreciating tools makes it difficult to identify dynamic parameters. In cases where the outputs have a lot of noise and in the structure that energy depreciating tools are used, blind source separation method is not accurate enough to estimate modal parameters. One of the important points in identifying the modal parameters is the type of external excitation of the structure. These excitations can be considered as stationary signals and nonstationary signals. Conventional structural identification methods are not able to identify modal parameters excited by nonstationary signals. In this dissertation, a method is proposed that by using the modified cross-correlation technique and performing operations on the cross-correlation of structural responses, it is possible to identify the modal parameters stimulated by the unstable signal. Also, the accuracy of calculating the damping ratio in the presence of medium to high damping has increased. To evaluate the performance of the proposed method, numerical simulation has been used which the results show that with the proposed method, natural frequencies and shape of structural modes with an error of less than 1% have been estimated. Damping ratios with less than 6% error were also obtained. A laboratory model has been used to validate presented method.