In this paper, a new locating method based on the optimization method for estimating the position of an electric dipole source in underwater environments with a plate uniform circular antenna (PUCA) is presented. The image principle is introduced to build the manifold of the PUCA, greatly reducing the complexity of the manifold. The evaluation function is obtained using the mixed polarization multiple signal classification algorithm, where the minimum value is found using the fast optimal method improved matrix adaptation evolution strategy (MA-ES). In this locating method, the voltage from each channel of the PUCA is the input spatial-temporal data. As a result, the 3-D field components are reduced, and the method can easily be implemented in practical engineering applications. The theoretical analysis and the experiments conducted for both the simulation and the actual received data demonstrate that the accuracy performance of the locating method based on the improved MA-ES is higher than that of the MA-ES and of the covariance MA-ES.