This study presents a hybrid optimization and decision-making framework for enhancing the performance of Electrical Discharge Machining (EDM) using graphite electrodes. Three key performance indicators—surface roughness (Ra), material removal rate (MRR), and electrode wear rate (EWR)—are simultaneously optimized using a multi-objective approach. To address the nonlinear nature of the process and the limited experimental data obtained from a Taguchi L18 design, Gaussian Process Regression (GPR) is employed as a surrogate modeling technique. The predicted models are used within the Non-dominated Sorting Genetic Algorithm II (NSGA-II) to generate a Pareto front of optimal process parameter configurations. To select the most suitable solution from the Pareto set, the Analytic Hierarchy Process (AHP) is applied, with consistency ratio (CR) analysis ensuring reliable pairwise judgments. For comparative validation, additional rankings are performed using three multi-criteria decision-making (MCDM) methods—TOPSIS, SAW, and MABAC—under Entropy and MEREC weighting schemes. Results reveal that the optimal configuration identified by AHP and MABAC (with Entropy weights) achieves the highest MRR (98.46 mg/min) while maintaining an acceptable surface quality (Ra = 3.15 µm) and electrode wear (EWR = 2.47 mg/min). The proposed NSGA-II–AHP framework provides a robust and flexible strategy for balancing trade-offs in EDM performance and can be extended to other manufacturing optimization problems with conflicting objectives.