We previously demonstrated that a high-fat diet (HFD) consumption can cause not only peripheral insulin resistance, but also neuronal insulin resistance. Moreover, the consumption of an HFD has been shown to cause mitochondrial dysfunction in both the skeletal muscle and liver. Rosiglitazone, a peroxizome proliferator-activated receptor-γ ligand, is a drug used to treat type 2 diabetes mellitus. Recent studies suggested that rosiglitazone can improve learning and memory in both human and animal models. However, the effects of rosiglitazone on neuronal insulin resistance and brain mitochondria after the HFD consumption have not yet been investigated. Therefore, we tested the hypothesis that rosiglitazone improves neuronal insulin resistance caused by a HFD via attenuating the dysfunction of neuronal insulin receptors and brain mitochondria. Rosiglitazone (5 mg/kg · d) was given for 14 d to rats that were fed with either a HFD or normal diet for 12 wk. After the 14^th week, all animals were euthanized, and their brains were removed and examined for insulin-induced long-term depression, neuronal insulin signaling, and brain mitochondrial function. We found that rosiglitazone significantly improved peripheral insulin resistance and insulin-induced long-term depression and increased neuronal Akt/PKB-ser phosphorylation in response to insulin. Furthermore, rosiglitazone prevented brain mitochondrial conformational changes and attenuated brain mitochondrial swelling, brain mitochondrial membrane potential changes, and brain mitochondrial ROS production. Our data suggest that neuronal insulin resistance and the impairment of brain mitochondria caused by a 12-wk HFD consumption can be reversed by rosiglitazone.