<p>This research reveals an innovative hybrid optimization strategy to achieve sustainable manufacturing in 3-axis CNC face milling operations. The proposed framework merges the Non-dominated Sorting Genetic Algorithm III (NSGA-III) with a Self-Adaptive strategy Generation Engine (SAGE) to simultaneously optimize four indicators: production time, production cost, energy consumption, and surface roughness. The cutting energy is defined as the multiplication of the net idle cutting power by the required time. The cutting power formulation is based on an industrial case study, where the Gaussian Process Regression (GPR) model, trained on experimental data, is employed to predict spindle power consumption as a function of cutting parameters. The goal of this paper is to optimize a four-objective problem, based on the novel hybrid method NSGA-III/SAGE, improving both local exploration and global search abilities during the optimization process. In order to validate the method, two Benchmark functions of DTLZ1-4D and DTLZ2-4D has been deployed. Five performance metrics: Hypervolume (HV), Spread (Δ), Spacing (S), Inverted Generational Distance (IGD), and Delta Metric (Δ′) are taken as evaluation indexes. The developed approach shows an excellent performance regarding tool metrics values. Consequently, NSGA-III/SAGE is adopted to optimize an industrial cutting process. Numerical results highlight superior convergence behavior and improved solution diversity with the complete SAGE configuration. Notably, disabling key SAGE modules resulted in a 28.57% increase in computation time, a 12.13% decrease in HV, a 32% rise in production time, and a 6% worsening of surface roughness. Additionally, performance degradation is observed in the Spread metric. These findings underscore the critical importance of integrating all SAGE components to effectively balance multiple conflicting objectives, thereby enabling intelligent and sustainable process planning.</p>