<p>Traditional trial-and-error methods are time-consuming and laborious, and it is difficult to find the optimal combination of element concentration to achieve a balance between high strength and high electrical conductivity in copper alloys. Therefore, an improved strategy for the reverse design of copper alloy composition is proposed in this paper. Specifically, a forward model (predicting properties from composition) and a reverse model (predicting composition from properties) between composition and properties were successively constructed in this paper. Then, by combining these two models and introducing a genetic algorithm to expand the potential composition, an improved alloy composition design system was established. Ultimately, the compositions of four alloys were successfully designed according to four target properties. The error between the predicted properties and the target properties is less than 5%, outperforming the existing design system. This method provides a reference for the reverse design of alloy materials.</p> Graphical abstract <p></p>