In this paper detection of epileptic seizure and non-seizure patient detected using the electroencephalogram (EEG). EEG of human brain is obtained by recording of EEG signal it is related to the medical field specialty, diagnosis of brain and brain related disease Proposed method based on wavelet coefficients types such as DWT, CWT, SWT (stationary wavelet transform) being translational invariant.in wavelet coefficient we use different properties. wavelet transform used for feature extraction and classifier used are support vector machine. Different artefact are removed using ICA.EEG advantageous for neurological disease if the uses of more feature extraction accuracy is better and also error rate is low.