Objective To investigate the therapeutic effect of andrographolide on brain injury caused by chronic hypoperfusion and its mechanism. Methods (1) Animal experiment. A total of 40 SD rats were assigned to sham operation group, modeling group, or low⁃, medium⁃, and high⁃dose groups according to the random number table method, with 8 rats in each group. Rats of the modeling and 3 different doses groups received permanent bilateral common carotid artery occlusion by bilateral common carotid artery occlusion method, while the sham operation group only underwent sham operation. After 28 days of modeling, the low⁃, medium⁃, and high⁃dose groups received intragastric administrations of andrographolide in 20 mg/kg, 40 mg/kg and 80 mg/kg, respectively, whereas the sham operation and modeling groups received intragastric administration of normal saline in equivalent volume, once a day, for 14 days continuously. (2) Cell experiment. Cells were assigned to control group, modeling group, 1 μmol/L andrographolide group, 2 μmol/L andrographolide group, or 4 μmol/L andrographolide group. Various andrographolide intervention groups received andrographolide in corresponding concentrations for intervention, while equivalent volume complete medium was added to the control and modeling groups. After 24⁃hour intervention, except for the control group, cells of the remaining 4 groups were placed in a closed anoxic chamber and the model was established by oxygen⁃glucose deprivation for 6 hours. (3) Ethology was detected and learning memory function of rats was evaluated by water maze and open field test. The pathological states of nerve cells in brain tissues were observed by performing the HE and Nissl staining on brain tissues of rats. The ELISA kit was employed to detect expressions of inflammatory factors in brain tissues. The CCK⁃8 and Hoechst staining were used to detect and observe HT22 cell active state. The Western blot was employed to detect expressions of cell and proteins related to nuclear factor kappa B (NF⁃κB) pathway in brain tissues of rats. Results (1) Animal experiment. Compared with the sham operation group, the modeling group exhibited longer escape latency and the first time to cross platform, whereas reduced platform crossing frequency, upright frequency, and reduced number of lattice traversal; in addition, compared with the modeling groups, rats of the 3⁃dose groups obtained shorter escape latency and the first time to cross platform, while increased platform crossing frequency (P<0.05). Neurons in hippocampus and cortical areas of rats in the modeling group presented as disordered arrangement, cytoplasmic atrophy, and nucleolar breakage. Among the 3⁃dose groups, neurons in hippocampus and cortical areas of rats yielded relieved injury degree as compared with the modeling group. Compared with the sham operation group, Nissl bodies in neurons in hippocampus and cortical areas of rats were significantly reduced in the modeling group, and neurons were impaired. Among the 3⁃dose groups, the number of Nissl bodies in neurons of rats increased, and the number of injured neurons was significantly decreased as compared with the modeling group. Compared with the sham operation group, the modeling group exhibited elevated expressions of interleukin (IL)⁃6, IL⁃1β, and tumor necrosis factor (TNF)⁃α, whereas a decreased IL⁃10 expression. Compared with the modeling group, the low⁃ and medium⁃dose groups yielded decreased expressions of IL⁃1β and TNF⁃α, while an elevated IL⁃10 expression. The medium⁃dose group interpreted decreased IL⁃6 expression, while the high⁃dose group depicted an elevated IL⁃10 expression (P<0.05). Compared with the sham operation group, the modeling group demonstrated increased expressions of phospho⁃IκBα and phospho⁃p65; moreover, compared with the modeling group, the 3⁃dose groups expressed decreased expressions of phospho⁃IκBα and phospho⁃p65, but the expression of phospho⁃p65 decreased while expression of phospho⁃IκBα increased successively with the increase of dose (P<0.05). (2) Cell experiment. Compared with the control group, cells of the modeling group presented as a decreased activity, while compared with the modeling group, cells in the 2 μmol/L and 4 μmol/L andrographolide groups yielded increased activities. Compared with the control group, cells of the modeling group obtained an increased protein expression of phospho⁃p65; furthermore, compared with the modeling group, cell phospho⁃p65 expression of the 3 andrographolide intervention groups was decreased (P<0.05). Conclusion Andrographolide can ameliorate cognitive impairment and autonomic motor disability of rats with hypoperfusion, relieve anxiety status, its mechanism may be related to inhibit activation of NF⁃κB pathway and regulate expressions of inflammatory factors.