A study was conducted between November 2006 and October 2009 to determine the factors predicting the presence and prognosis of epilepsy in patients with cerebral palsy. We enrolled 2 groups of patients: 42 with cerebral palsy in group 1 and 56 patients with cerebral palsy and epilepsy in group 2. The subjects in group 2 were considered to have good epilepsy prognosis if they were free of seizures for the previous year; otherwise they were considered to have poor epilepsy prognosis. In group 2, neonatal epilepsy, family history of epilepsy, and moderate to severe mental retardation were significantly higher than in group 1 (P < 0.05). In univariate analysis, neonatal seizures, epileptic activity as measured by electroencephalography, and polytherapy were found to be predictors of poor epilepsy prognosis. Additionally, the need for long-term medication to control seizures unfavorably affects prognosis. In logistic regression analysis, neonatal seizure and interictal epileptic activity in electroencephalography were found to be independent predictors of poor epilepsy outcome. In addition, logistic regression analysis revealed that increasing age reduces the success of epilepsy treatment. Neonatal seizures, family history of epilepsy, and mental retardation were found to be important and independent predictors of development of epilepsy in patients with cerebral palsy.