Glutathione peroxidase 4 (GPX4) is distinguished from other members of the GPX family as being the enzyme capable of reducing phospholipid hydroperoxides within cellular membranes and therefore protecting cells from ferroptosis, a form of iron-driven cell death involving lipid peroxidation. We previously identified a homozygous point mutation in the GPX4 gene, resulting in an R152H coding mutation and a substantial loss of GPX4 enzymatic activity, in patients with Sedaghatian-type spondylometaphyseal dysplasia (SSMD), an ultrarare progressive disorder. To explore whether selective binding and correction of the loss of enzyme activity observed with this variant is possible, we screened 2.8 billion compounds in a DNA-encoded chemical library and identified compounds with remarkably selective binding affinities with the R152H variant (GPX4R152H) over wild-type (GPX4WT). Our structural optimization of these compounds led to the identification of analogues with improved potency for R152H GPX4. The most promising compounds selectively restored the enzyme activity of GPX4R152H and specifically increased the viability of fibroblast and lymphoblast cells developed from an SSMD patient with the homozygous R152H variation but not control cells from a healthy parent or HEK293T cells undergoing ferroptosis induced by a wild-type GPX4 inhibitor. This approach represents a low-cost, high-throughput, and generalizable approach to identify targeted small-molecule therapeutics for missense variants, which features the potential to be broadly applied to diseases that bear point mutations on crucial proteins, including cancers.