Smooth muscle cell (SMC) phenotypic switch from a quiescent ‘contractile’ phenotype to a dedifferentiated and proliferative state underlies the development of cardiovascular diseases (CVDs); however, our understanding of the mechanism is still incomplete. In the present study, we explored the potential role of ferroptosis, a novel nonapoptotic form of cell death, in SMC phenotypic switch and related neointimal formation. We found that ferroptotic stress was triggered in cultured dedifferentiated SMCs and arterial neointimal tissue of wire-injured mice. Moreover, pro-ferroptosis stress was activated in arterial neointimal tissue of clinical patients who underwent carotid endarterectomy. Blockade of ferroptotic stress via administration of a pharmacological inhibitor or by global genetic overexpression of glutathione peroxidase-4 (GPX4), a well-established anti-ferroptosis molecule, delayed SMC phenotype switch and arterial remodelling. Conditional SMC-specific gene delivery of GPX4 using adreno-associated virus in the carotid artery inhibited ferroptosis and prevented neointimal formation. Conversely, ferroptosis stress directly triggered dedifferentiation of SMCs. Transcriptomics analysis demonstrated that inhibition of ferroptotic stress mainly targets the mitochondrial respiratory chain and oxidative phosphorylation. Mechanistically, ferroptosis inhibition corrected the disrupted mitochondrial homeostasis in dedifferentiated SMCs, including enhanced mitochondrial ROS production, dysregulated mitochondrial dynamics, and mitochondrial hyperpolarization, and ultimately inhibited SMC phenotypic switch and growth. Copper-diacetyl-bis-methylthiosemicarbazone (CuATSM), an agent used for clinical molecular imaging and that potently inhibits ferroptosis, prevented SMC phenotypic switch, neointimal formation and arterial inflammation in mice. These results indicate that pro-ferroptosis stress is likely to promote SMC phenotypic switch during neointimal formation and imply that inhibition of ferroptotic stress may be a promising translational approach to treat CVDs with SMC phenotype switch.