Epilepsy is a chronic neurological disorder characterized by sudden occurrence of excessive neuronal discharges which affects most of the people. Epileptic patients are suffering from seizures which cause damage to the neural tissues. It also results in many injuries such as fractures, burns, accident and death. Many methods have been developed for seizure prediction. These methods extract various features from EEG signal and train the classifier to find the seizure appearance. Selecting effective features is very important for seizure detection. The features are obtained using exhaustive spectral and statistical analysis. Signal classification will be done using Artificial Intelligent classifiers. The proposed device is easy to wear and feel comfortable long period of time. A wearable device will be designed which would contain minimum number of electrodes to accurately predict and detect the occurrence of seizure. The signal processing unit will be miniaturized and be made like a pocket device so that it would not affect the social stigma of the patient. This device would be used for chronic epileptic patients where the occurrence of seizure is more frequent.  It can predict the occurrence of seizure with accuracy and less number of false positive rate