Mongolian gerbils (28 males) were divided into four groups, control (C), treadmill-exercised (Ex), treated with the extract of Diospyros kaki (Dk), and exercised plus treated with the Dk extract (Ex+Dk). Animals of the respective groups were running-exercised for 30 min per day during 8 weeks, and the Dk extract (dose 20 mg/kg) was given by gavage during five days per week within the same period. After the treatment and exercise period, an epilepsy model was produced by penicillin G injection (500 IU) into the left somatomotor cortex, and the electrocorticogram (ECoG) was recorded during 120 min. The mean frequency of spike/wave complexes was significantly smaller in the Ex and Ex+Dk groups from the 65th min of the observation period and, in the Dk group, from the 75th min than the respective value in the C group (P < 0.01, P < 0.05, and P < 0.01, respectively). The differences in the amplitude values and latency to onset of the spike/wave events among all groups did not reach the significance level (P > 0.05). Thus, both the running exercise and Dk extract applications inhibit penicillin-induced epileptiform activity by altering the spike/wave frequency or severity of seizures observed in ECoG recordings. Further studies are needed to determine the effects of physical activity of different intensities and forms and to analyze the active compounds in the Dk extract.