This literature review explores advancements in Prediction of Epileptic Seizures with Machine Learning model and Deep Learning techniques. The unpredictability of epileptic seizures serious difficulties to patient safety and quality of life recent the research makes use of EEG-based feature extraction and classification. The models and hybrid deep learning architectures recognize states Traditional machine learning approaches, such as SVM. Have worked well with engineered features, Random Forest CNN and LSTM models can reach more accurate results by learning create sophisticated rhythmic and color designs from EEG data. Important artifacts still remain despite some removal activity imbalanced dataset, personalization, and real-time deployment. This key methodologies, comparative performance, review highlights Interpretation and progress aimed at creating sturdy and practical seizure prediction systems.