Abstract: The eigenvector method for maximum-likelihood estimation of phase error can obtain ideal performance of phase error estimation by using the eigenvector corresponding to its largest eigenvalue. Although the method is accurate and robust, it requires eigen-decomposition of the sample covariance matrix, which is computationally expensive and limits its real-time applications. In this paper, a Weighted Maximum Norm Method (WMNM) for phase error estimation is proposed. The eigenvector of the maximum eigenvalue can be obtained directly by solving the problem of maximizing L-2 norm, which avoids the eigen-decomposition of the sample covariance matrix and reduces the computational cost greatly. By adding different weights to each range bin, the contribution of the range cells with high SNR can be enhanced. Experimental results of the measured data by SAR and Inverse SAR (ISAR) verify the validity of the proposed algorithm.