The archetype inhibitors of ferroptosis, ferrostatin-1 and liproxstatin-1, were identified via high-throughput screening of compound libraries for cytoprotective activity. These compounds have been shown to inhibit ferroptosis by suppressing propagation of lipid peroxidation, the radical chain reaction that drives cell death. Herein, we present the first rational design and optimization of ferroptosis inhibitors targeting this mechanism of action. Engaging the most potent radical-trapping antioxidant (RTA) scaffold known (phenoxazine, PNX), and its less reactive chalcogen cousin (phenothiazine, PTZ), we explored structure-reactivity-potency relationships to elucidate the intrinsic and extrinsic limitations of this approach. The results delineate the roles of inherent RTA activity, H-bonding interactions with phospholipid headgroups, and lipid solubility in determining activity/potency. We show that modifications which increase inherent RTA activity beyond that of the parent compounds do not substantially improve RTA kinetics in phospholipids or potency in cells, while modifications that decrease intrinsic RTA activity lead to corresponding erosions to both. The apparent "plateau" of RTA activity in phospholipid bilayers (kinh ∼ 2 × 105 M-1 s-1) and cell potency (EC50 ∼ 4 nM) may be the result of diffusion-controlled reactivity between the RTA and lipid-peroxyl radicals and/or the potential limitations on RTA turnover/regeneration by endogenous reductants. The metabolic stability of selected derivatives was assessed to identify a candidate for in vivo experimentation as a proof-of-concept. This PNX-derivative demonstrated stability in mouse liver microsomes comparable to liproxstatin-1 and was successfully used to suppress acute renal failure in mice brought on by tissue-specific inactivation of the ferroptosis regulator GPX4.