Abstract: To address the high trial-and-error cost and the difficulty in synergistically optimizing multiple properties within a wide composition space of the Cu-Al-Ni-Fe-Mn quinary copper alloy system, this study developed a composition design framework for copper alloys based on the differential evolution algorithm. A total of 214 sample data were collected by integrating literature data and industrial data. A 117-dimensional feature vector was constructed based on 58 elemental physicochemical properties and discrete encoding of process states, and the feature dimension was reduced through a feature selection strategy. Weighted ensemble prediction models combining random forest and multilayer perceptron algorithms were established separately for three target properties, including ultimate tensile strength, elongation, and corrosion rate. The differential evolution algorithm was embedded into the ensemble models as an evaluator of the comprehensive objective function to minimize the normalized weighted deviation of multiple performances in the constrained composition space. The results show that under the target conditions of ultimate tensile strength no less than 690 MPa, elongation no less than 17%, corrosion rate no more than 0.011 mm/a, and annealed process state, the algorithm achieves stable convergence at approximately the 60th generation. The optimized composition in mass fraction is determined as Cu 80.2%, Al 9.70%, Ni 4.78%, Fe 2.91%, Mn 2.14%, Sn 0.055%, Si 0.134%, Zn 0.067%, and Pb 0.015%. The ensemble model predicts that the corresponding ultimate tensile strength, elongation, and corrosion rate reach 690.000 2 MPa, 17.5%, and 0.014 4 mm/a, respectively. Ingots prepared with the optimized composition were annealed at 675 ℃ for 2 h. The measured average values of ultimate tensile strength, elongation, and corrosion rate are 689 MPa, 16.75%, and 0.012 mm/a, respectively, which fully satisfy the specified performance requirements. The deviations of ultimate tensile strength and elongation are both within 2%. The proposed framework exhibits engineering applicability for the multi-objective synergistic optimization of copper alloys, which provides a feasible technical approach for the intelligent composition design of high-performance copper alloys.