System parameter identification can be divided into two research directions: time-invariant system parameter identification and time-varying system parameter identification. The research on time-invariant system parameter identification has become mature, while the research on time-varying system parameter identification is still in its infancy. For multi-degree-of-freedom time-varying structures, a method for parameter identification based on time frequency slice decomposition is proposed in this paper. The first step of this method is to collect the displacement response of the structure, and then obtains the time-frequency energy distribution of the response in the whole time-frequency band. According to the time-frequency distribution characteristics of the structure, multiple time-frequency window signals are sliced out, and the inverse transform of FSWT is performed to reconstruct the signals in time domain. The reconstructed signal corresponds to each order modal displacement response signal of the structure. The instantaneous frequency of the signal is extracted by Hilbert transform, and the modal frequencies of each order of the structure are identified. According to a simulation experiment of the three-degree-of-freedom mass spring damping system, the method is proved to have good identification accuracy and practical engineering value.