For several decades, it has been recognized that overactivation of the glutamate-gated N-methyl-D-aspartate receptors (NMDARs) and subsequent Ca2+ toxicity play a critical role in ischemic brain injury. 24S-hydroxycholesterol (24S-HC) is a major cholesterol metabolite in the brain, which has been identified as a potent positive allosteric modulator of NMDAR in rat hippocampal neurons. We hypothesize that 24S-HC worsens ischemic brain injury via its potentiation of the NMDAR, and reducing the production of 24S-HC by targeting its synthetic enzyme CYP46A1 provides neuroprotection. We tested this hypothesis using electrophysiological, pharmacological, and transgenic approaches and in vitro and in vivo cerebral ischemia models. Our data show that 24S-HC potentiates NMDAR activation in primary cultured mouse cortical neurons in a concentration-dependent manner. At 10 µmol/L, it dramatically increases the steady-state currents by 51% and slightly increases the peak currents by 20%. Furthermore, 24S-HC increases NMDA and oxygen-glucose deprivation-induced cortical neuronal injury. The increased neuronal injury is largely abolished by NMDAR channel blocker MK-801, suggesting an NMDAR-dependent mechanism. Pharmacological inhibition of CYP46A1 by voriconazole or gene knockout of Cyp46a1 dramatically reduces ischemic brain injury. These results identify a new mechanism and signaling cascade that critically impacts stroke outcome: CYP46A1 → 24S-HC → NMDAR → ischemic brain injury. They offer proof of principle for further development of new strategies for stroke intervention by targeting CYP46A1 or its metabolite 24S-HC.