In this paper, a set of inversion algorithms for the micro-motion parameters in a retroreflector array was established based on the time-frequency signals of laser dynamic speckles. Firstly, the real-time intensitys formula of laser speckle from a retroreflector and a retroreflector array were derived by a physical optics method. Then, based on the short-time Fourier transform, the formation mechanism and digital characteristics of the speckle power spectrum were studied. Finally, the spectral correlation method and time-frequency to phase-amplitude transform algorithm were proposed, and the algorithm was used to extract the period and amplitude distribution of the time-frequency spectral lines of dynamic speckles. Besides, we made an inversion on the spin period and the direction of the rotation axis of a target in three typical motion states. The results show that the proposed inversion algorithm can obtain the rotation period and line-of-sight angle with high accuracy through speckle intensity signals in a few periods, but the inversion accuracy for the azimuth angle of the rotation axis is relatively poor. It follows that more observation data are required for satisfactory results.