The radar cross section, high-resolution range profile (HRRP) and micro-Doppler based feature extraction algorithm may be invalid in the case of small changes line-of-sight (LOS) targets with low radar pulse repetition frequency. A micro-motion period estimate algorithm based on the periodic matching feature of inverse synthetic aperture radar (ISAR) image sequence is proposed. Based on the analysis of ISAR imaging of micro-motion targets, we use the similarity between the images of ISAR sequence to estimate the micromotion period. By using the minimum-entropy based criterion to select the base-image with linear LOS change and the methods of dynamic range compression and full image matching based similarity estimation, the algorithm achieves high stability and low computational complexity. And the algorithm has low constraints for the radar data rate and ISAR imaging quality which is suitable for engineering accomplishment. The simulation shows the validity and high stability of the estimation algorithm in the wobble and precession cases. The influence of the base-image selection and signal-to-noise ratio is also analyzed.