This paper presents a framework for Electroencephalography (EEG) seizure prediction in time domain. Moreover, it studies an efficient lossy EEG signal compression technique and its effect on further processing for seizure prediction in a realistic signal acquisition and compression scenario. Compression of EEG signals are one of the most important solutions in saving speed up signals transfer, reduction of energy transmission and the required memory for storage in addition to reduction costs for storage hardware and network bandwidth. The main objective of this research is to use trigonometric compression techniques including; Discrete Cosine Transform (DCT) and Discrete Sine Transform (DST) algorithms on EEG signals and study the impact of the reconstructed EEG signals on its seizure prediction ability. Simulation results show that the DCT achieves the best prediction results compared with DST technique achieving sensitivity of 95.238% and 85.714% respectively. The proposed approach gives longer prediction times compared to traditional EEG seizure prediction approaches. Therefore, it will help specialists for the prediction of epileptic seizure as earlier as possible.