Epilepsy is a spontaneous or provoked synchronized cortical activity in cerebral neurons that leads to epileptic seizures which affect the person in motorized dispositions and sensory or psychological sensations. The Detection of epilepsy is achieved using the electroencephalogram (EEG). The EEG is a continuous, irregular signal in which any section is never repeated exactly at any other time. The EEG can therefore not be exactly described by certain parameters (such as amplitude or frequency) It is only available for statistical estimation. With the properties mentioned, the EEG virtually challenges signal analysis. Frequency analysis (spectral analysis) is an essential component of electronic signal processing in many devices in EEG diagnostics. In general, the digital representation and processing of the EEG is becoming increasingly important. This paper presents a procedure to detect seizures in EEG signal using EEG wave component frequency analysis. Depending on the level of alertness, a distinction is made between different patterns in the electroencephalogram, namely Alpha, Beta, Theta and Delta waves. The statistical parameters are extracted after wavelet transform is applied on the EEG signal. These parameters are trained with SVM classifier to detect the seizure. The proposed method performs better than the existing algorithms.