Abstract <p>The article considers the fundamentals of control of the structure of new metals based on iron, aluminum, etc. synthesized by melting using conventional technologies and technologies that employ quantum-mechanical calculations of the alloying chemical element. Optimal compositions, average grain sizes, and structural stability of new metals are determined using neural networks, employing fractal analysis and nonlinear dynamics. A digital twin of the target metal is developed, including a series of twins. The results of predicting the structure and mechanical properties of new materials according to the developed methodology are presented.</p>