Trigeminal neuralgia (TN) is a severe chronic neuropathic pain that mainly affects the distribution area of the trigeminal nerve with limited treating efficacy. There are numerous treatments for TN, but currently the main clinical approach is to suppress pain by carbamazepine (CBZ). Brain-derived neurotrophic factor (BDNF) is closely related to chronic pain. This study aims to determine the effects of CBZ treatment on BDNF expression in both the trigeminal ganglion (TG) and serum of TN via a chronic constriction injury of the infraorbital nerve (ION-CCI) rat model. The ION-CCI models were established in male Sprague-Dawley rats and were randomly divided into a sham group, a TN group, a TN+low-dose CBZ treatment group (TN+20 mg/kg CBZ group), a TN+medium-dose CBZ treatment group (TN+40 mg/kg CBZ group), and a TN+high-dose CBZ treatment group (TN+80 mg/kg CBZ group). The mechanical pain threshold in each group of rats was measured regularly before and after surgery. The expressions of BDNF and tyrosine kinase receptor B (TrkB) mRNA in TGs of rats in different groups were determined by real-time PCR, and the expression of BDNF protein on neurons in TGs was observed by immunofluorescence. Western Blotting was used to detect the protein expression of BDNF, TrkB, extracellular regulated protein kinases (ERK), and phospho-extracellular regulated protein kinases (p-ERK) in TGs of rats in different groups. The expression of BDNF in the serum of rats in different groups was detected by enzyme-linked immunosorbent assay (ELISA). The results of mechanical pain sensitivity showed that there was no significant difference in the mechanical pain threshold in the right facial sensory area of the experimental rats in each group before surgery (all P>0.05). From the 3rd day after operation, the mechanical pain threshold of rats in the TN group was significantly lower than that in the sham group (all P<0.01), and the mechanical pain threshold of rats in the TN+80 mg/kg CBZ group, the TN+40 mg/kg CBZ group, and the TN+20 CBZ mg/kg group was higher than that in the TN group (all P<0.05). The BDNF and TrkB mRNA and protein expressions in TGs of rats in the TN group were higher than those in the sham group (all P<0.05), and those in the TN+80 mg/kg CBZ group, the TN+40 mg/kg CBZ group, and the TN+20 mg/kg CBZ group were lower than the TN group (all P<0.05). The p-ERK levels in TG of rats in the TN+80 mg/kg CBZ group, the TN+40 mg/kg CBZ group, and the TN+20 mg/kg CBZ group were significantly decreased compared with the TN group (all P<0.05). The BDNF and neuron-specific nuclear protein (NeuN) were mainly co-expressed in neuron of TGs in the TN group and they were significantly higher than those in the sham group (all P<0.05). The co-labeled expressions of BDNF and NeuN in TGs of the TN+ 80 mg/kg CBZ group, the TN+40 mg/kg CBZ group, and the TN+20 mg/kg CBZ group were lower than those in the TN group (all P<0.05). The results of ELISA showed that the level of BDNF in the serum of the TN group was significantly higher than that in the sham group (P<0.05). The levels of BDNF in the TN+80 mg/kg CBZ group, the TN+40 mg/kg CBZ group, and the TN+20 mg/kg CBZ group were lower than those in the TN group (all P<0.05). Spearman correlation analysis showed that the BDNF level in serum was negatively correlated with mechanical pain threshold (r=-0.650, P<0.01). CBZ treatment can inhibit the expression of BDNF and TrkB in the TGs of TN rats, reduce the level of BDNF in serum of TN rats and the phosphorylation of ERK signaling pathway, so as to inhibit TN. The serum level of BDNF can be considered as an indicator for the diagnosis and prognosis of TN.