Optimising the parameters of leaching while assessing the dynamics of the process kinetics requires an investigation of the effects of different variables. A neural network (NN) and a fuzzy inference system (FIS) were used to evaluate and examine the leaching process. The results showed that increasing the acid concentration and stirring speed, while decreasing the solid-to-solution ratio and pH, enhanced copper (II) leaching. The optimum values obtained from the leaching process for pH 3 were found to be a solid-to-liquid ratio of 1g/100 mL, an agitation speed of 300 rpm, and an acid concentration of 1 M, with a 97% recovery of copper (II). Diffusion throughout the product layer controlled the leaching rate, and the experimental results suggested that a diffusion-controlled model would provide the best fit. The diffusion-controlled mechanism was indicated by an activation energy of 16.01 kJ/mol. To optimise the parameters of the leaching process, the algorithm training for neural networks (NNs) included the Levenberg–Marquardt method with a membership structure of 7-7-7-7, using the backpropagation (BP) technique for learning. The neural network (NN) method was trained using four input variables, representing leaching parameters, fifteen hidden layers, and one output representing copper (II) leaching recovery. R² values of 0.996, 0.997, and 0.997, respectively, show the validation, testing, and training phases of the ideal trained neural network. An R² value of 0.999 for FIS indicates that the study data can be precisely predicted. ANFIS had a Pearson's chi-squared value of 0.225, surpassing the ANN's score of 0.658.