Exposure to sevoflurane and other inhalational anesthetics can induce neurodegeneration in the developing brain. Although dexmedetomidine (DEX) has provided neuroprotection against hypoxic ischemic injury, relatively little is known about whether it has the neuroprotective effects against anesthetic-induced neurodegeneration. This study examined whether DEX improves the long-term cognitive dysfunction observed after exposure of neonatal rats to 3% sevoflurane. Seven-day-old rats received intraperitoneal saline (DEX 0) or DEX (6.6, 12.5, 25 μg/kg) 30 min before exposure to 3% sevoflurane with 21% oxygen for 4 h (n = 10 per group). The pups in the control group received only DEX 25 μg/kg without anesthesia. The escape latency in the Morris water maze was significantly increased in the DEX 0 group compared with the sham and control group, and the escape latency, but not the swimming path length, was significantly shorter at post-natal day 47 in the DEX 25 than in the DEX 0 group. The percent time spent in the quadrant was significantly decreased in the DEX 0 group compared with the sham and control group, and the percent time spent in the quadrant was significantly increased in the DEX 25 group compared with the DEX 0 groups. The freezing times of the DEX 0 and 6.6 groups were significantly decreased compared with those in the sham, control and DEX 25 groups. The number of NeuN-positive cells in the CA1 region was significantly decreased in the DEX 0 and 6.6 groups compared with the sham, control and DEX 25 groups. These findings indicate pre-treatment with DEX may improve long-term cognitive function and ameliorate the neuronal degeneration induced by sevoflurane exposure in neonatal rats.