Epilepsy is a neurological disorder induced by brain neurons’ spontaneous activation and discharge. Neurophysiologists’ manual identification of seizures and their complex wave forms in electroencephalography (EEG) is time-consuming, laborious, and prone to human error. The proposed model introduced an automated epilepsy seizure detection system utilizing a continuous wavelet transform-based pre-processed deep ensemble neural network. The EEG signals are converted into time- frequency image representations, termed scalograms, which are then fed as input to the ensemble model containing Dense net 121 and Efficient Net B2. The fused features are then passed through the global average pooling layer followed by the soft maxlayer, which yields the multi-class classification of pre-ictal, inter-ictal and ictal classes. The proposed model ensured no misclassification cases by obtaining 100% accuracy, recall, precision, F1 score, and AUC score. The proposed model also outperformed the other pre-trained networks and the state-of-the-art techniques. Thus, the proposed model is ideal for automatic epilepsy seizure election and can aid neurologists in early diagnosis.