A sudden and excessive electrical activity of neurons in the brain is termed as epilepsy. A patient who is suffering from epilepsy has attacks when excitability of neurons inside the brain exceeds certain threshold value. Current epilepsy detection techniques include SPECT, PET, MRI, and continuous EEG monitoring. Electroencephalography (EEG) measures the brain‟s electrical activity directly by using electrodes which are attached to the scalp of the patient, while other methods record changes in metabolic activity or blood flow, which are indirect markers of brain electrical activity. Also these methods are costly and time consuming. EEG provides valuable information in the examination of patients suffering from epilepsy. The inter-ictal and ictal epileptiform discharges in the EEG are significantly different from the normal and background EEG and can be used as time markers for the onset of seizure. This paper proposes seizure detection technique using wavelet transform and Back Propagation Neural Networks (BPNN). The former is used for EEG decomposition and feature extraction while the latter is used for classification. Seizure data was fed to train the network for back propagation of error and the trained network was further assessed by testing it with new raw seizure data. The seizure and non-seizure data was classified with the accuracy of 97. 3%. © Research India Publications.