In this study, we studied the localization problem of weakly electric fish using their electric field activity. These fish emit electric fields and employ electroreceptors to understand their surroundings. The goal of this study is to localize individual fish within a 1.5-meter by 1.5-meter testing aquarium using a custom-built data acquisition system. This system incorporates a 3 by 3 silver electrode array, alongside an analog circuitry that effectively filters and amplifies the recorded electric field activity leveraging instrumentation amplifiers. Furthermore, the digitization and logging of these electric field measurements are facilitated through the utilization of FPGAs. In parallel, we capture the kinematic movements of the fish using a camera system positioned above the experimental setup. The precise tracking of fish motion is achieved through a custom image processing algorithm. This approach furnishes a groundtruth reference for the three-dimensional coordinates of the fish within testing aquarium. Consequently, the fish position data is synchronized with the corresponding recordings of electric field activity obtained from the nine electrodes. For the refinement of position estimation, we implemented an artificial neural network (ANN) architecture, incorporating nine input nodes (representing electrode recordings) and two output nodes (cartesian fish coordinates). Furthermore, a two-step electrode recording history is incorporated into the ANN framework to augment the precision of the position estimates. Comparative evaluations against a contemporary particle filtering technique manifest the improved predictive performance of our ANN-based methodology in leveraging electric field recordings for fish localization, surpassing the capabilities documented in extant literature.