This chapter develops a computationally efficient model for automatic patient-specific seizure prediction using a two-layer LSTM from multichannel intracranial electroencephalogram time-series data. We decrease the number of parameters by employing a smaller input size and fewer electrodes, thereby making the model a viable option for wearable and implantable devices. The prediction model is evaluated using 26 patients from the European iEEG dataset, which is among the largest available datasets for epileptic seizure. An automatic preprocessing technique also is applied based on a common average reference to remove artifacts from this dataset. The simulation results show that the model with its simple structure in conjunction with the mean postprocessing procedure performed the best, with an average AUC of 0.885. In this chapter also the effect of the seizure type on the system performance is analyzed, and results demonstrate that the seizure type has a considerable impact.