Abstract: The mutual coupling and amplitude-phase error of an array have great negative impact on the direction finding performance of the MUSIC algorithm,hence the paper addresses the problem of calibrating the array errors arising from mutual coupling and amplitude-phase error.Aiming at the uniform linear array,the paper presents an improved array calibration algorithm,which estimates the array errors matrix from a set of measured steering vectors obtained by eigenvalue decomposition.The improved algorithm fully utilizes the special structure of mutual coupling matrix of the uniform linear array,and it realizes the optimized calibration for array errors by alternating iterative solution.Simulation results demonstrate that the improved algorithm can give highly accurate parameter estimation and can be extended to the situation that the calibration sources azimuths have deviations.