Lithiation reactions driven by chemical potential offer a promising avenue for directly regenerating degraded lithium iron phosphate (LFP). However, the choice of solution system significantly influences the lithium supplementation where improper selection may result in poor lithium recovery or extremely slow kinetics. Herein, it is identified that the most critical factor affecting solution repair effectiveness is the redox potential of the anions in the solution, which determines whether spent LFP (SLFP) can undergo spontaneous lithiation under ambient conditions. Then, machine learning (ML) is used for prediction and screening of huge potential solution systems, and finally a general strategy is proposed: creating a low redox potential solution system that incorporates anions with either low redox potential or moderate redox potential at high concentrations. As a demonstration, the regenerated LFP by ascorbic acid and LiOH solution systems exhibits a high discharge capacity of 144 mAh g-1 at 1 C, retaining 96% of its capacity after 500 cycles at 5 C. This work establishes an important criteria for designing solution systems to restore degraded LFP, marking a significant advancement in the direct regeneration of cathode materials from spent lithium-ion batteries (LIBs).