The GH4175 alloy is a typical new difficult-to-deform superalloy. Optimizing its composition to enhance its deformation capability in the high-temperature single-phase region is a crucial prerequisite for preventing cracking during the cogging process of this alloy. By combining thermodynamic calculations, high-temperature tensile experiments, and machine learning methods, the composition of GH4175 was optimized. The key compositional elements that influence the γ′ phase volume fraction at 800 ℃, the γ′ phase dissolution temperature, and the alloy melting temperature are identified via design space screening and adaptive learning strategies. A relationship model between the content of these elements and high-temperature elongation is established, clarifying the compositional range that ensures excellent high-temperature ductility while maintaining a reasonable processing window and γ′ phase volume fraction at 800 ℃.