The constant technological progress in the metal-mechanical sector and the growing demand for the manufacturing of high-strength materials with complex morphologies have driven the development of non-traditional methods to make the production of metal components more efficient. Pulsed Electrochemical Machining (PECM) is a non-conventional technique that allows metal removal on numerous materials. This method uses the electrolysis process, so it is capable of machining without generating heat, chips, or tool wear, making it an exceptional technique with applications in sectors such as aerospace, medical, and tool manufacturing. However, understanding its operation is challenging due to the chemical and electrical principles involved. In this sense, the present research develops an adaptive neuro-fuzzy inference system (ANFIS) optimized with a genetic algorithm to model the material removal rate (MRR) and overcut (OC) using the process parameters with the highest correlation to the variables of the study. The model was trained in a supervised manner using a data set based on machining with AISI-1018 steel. After optimization, a coefficient of determination R2 of 0.94 for OC and 0.92 for MRR was obtained. Finally, the repeatability of these results is validated by obtaining a coefficient of variation of 4.8%, demonstrating precision and higher performance when ANFIS is tuned using a stochastic method, as opposed to when it is tuned using gradient descent-based techniques.