Poststroke epilepsy is a common problem in clinical practice. This study aimed to determine the predictors of poststroke epilepsy and to identify the risk factors.A total of 436 patients were enrolled for this study. Baseline characteristics were recorded. Patients were followed up for 3 years. According to the occurrence of late seizures (i.e., seizure occurring 8 days or later after the stroke), the patients were separated into two groups: a seizure group and a seizure-free group. Baseline characteristics were compared between the groups. Univariate and multivariate Cox regression analysis was used to identify the risk factors for occurrence of poststroke epilepsy. The study was registered in the University Hospital Medical Information Network Center Clinical Trials Registry (registration number UMIN000009659, date 28/December/2012).Among the 436 patients, 26 developed late seizures-an incidence rate of 5.96%. In univariate analysis, female, subarachnoid hemorrhage, hemorrhagic lesion, lesion affecting the cortex, lesion affecting the frontal lobe, and absence of comorbidities were the significant risk factors for occurrence of epilepsy. In multivariate analysis, the independent predictors of poststroke epilepsy were hemorrhagic lesion (hazard ratio (HR) = 3.03) and lesion location in the cortex (HR = 4.64). The incidence of poststroke epilepsy in patients with both two risk factors was 15.4%.Poststroke epilepsy occurs in almost 6% of patients within 3 years after stroke onset. Hemorrhagic stroke and subcortical lesion appear to be significantly associated with likelihood of developing epilepsy. Thus, prophylactic administration of antiepileptic drugs could be useful for stroke patients with these two risk factors.