The least-square migration methods can make the seismic profile imaging quality higher,and are conducive to better imaging for lithologic reservoir and reservoir parameter inversion.However,in the process of practical application,it is very difficult for least square migration to achieve high accuracy of lithology imaging.Based on an analysis of the problems existing in the traditional least squares migration,the authors have developed a least squares migration method based on phase matching,which abandons the quadratic function used frequently by the traditional methods,and employs the cross-correlation function;in this way,this method emphasizes more on the role of the seismic phase information in the least squares inversion imaging;in theory,and it can be better applied to real data.Based on the cross-correlation function building,synthesizing prediction data by reverse time demigration,gradient precondition,and pre-conditional conjugate gradient,the authors explored and realized a cross-correlation objective function for least-squares reverse migration.Simple sub-sag model and complex salt model tests show that the proposed method has applicability and effectiveness.