The effectiveness of neural-network modeling for the multi-objective optimization of the composition and the prediction of the strength of system-alloyed Fe–Cr–Ni–Mo-based maraging steels is shown. Analyzing the dependence of the mechanical characteristics of this class of steels on their composition, it is possible to determine the effect of small additives of alloying elements on yield stress, strength, toughness, and static crack resistance. For multi-objective optimization of the composition of maraging steels, the ideal-point method and the linear-sequence algorithm are used.