Abstract: An element coordinate estimating method for correlation sonar system is presented, which can be used to improve the performance of velocity estimation. The characteristics of correlation values around the maximum of the amplitude of the correlation function at delay zero are analyzed thoroughly. Localized Least Mean Squares (LLMS) criterion is put forward for element coordinates estimation. Sequential Quadratic Programming (SQP) is adopted as the optimization method. So the systematic signal processing method of correlation sonar for element coordinate estimation is established. The results of experimental data show that the algorithm can be used to accurately estimate element coordinate of correlation sonar. The estimated results help improve velocity measure precision of correlation sonar by 67%.