This investigation looked into the electrochemical machining of EN31 steel. To do this, we initially used neural networks to make predictions about the permeation flux and the fouling resistance (ANNs). Applied voltage, flow rate of electrolyte and feed rate of tool were among the experimental input data utilised to train ANN models. The modelling findings indicated a high degree of agreement between the experimental data and the predicted values. An optimization strategy was developed using a genetic algorithm to foresee the optimal operating settings for achieving the desired material removal rate. Comparison of predicted values with observed values demonstrates the model's reliability. © 2025 Elsevier B.V., All rights reserved.