In ISAR imagingthe phase error caused by nonuniform rotation is relevant to the radial position of scatterersso it is difficult for us to use traditional phase focusing algorithms to complete the phase compensationin which the uniform phase correction function is applied.In order to solve the problemwe put forward a phase compensation method based on adaptive Gaussian Chirplet decomposition (AGCD) fast algorithm.In this methodthe linear frequency modulation (LFM) signal parameters of two stronger scatterers in random range bin are obtained via AGCD algorithmand then the multiple prominent point processing (PPP) model is implemented to eliminate the phase error caused by translational motion and nonuniform rotation.This method has the advantages of low computation cost and high accuracyand it doesnt require the scatterer be well isolated.The numerical simulation proves its good efficiency.