This paper proposes a system for predicting epileptic seizures from EEG signals using Machine Learning approaches in order to prevent seizures through medication. Electrocorticography (ECoG) and electroencephalography (EEG) media are frequently used to detect these brain impulses. These signals generate a large amount of data and are complicated, noisy, non-linear, and non-stationary. Therefore, identifying seizures and learning about the brain's functions is a difficult undertaking. Without sacrificing performance, machine learning classifiers can classify EEG data, detect seizures, and highlight pertinent, meaningful patterns. In this study, the epileptic seizure dataset was classified using a variety of classifiers. Support vector machines performed better than Naive Bayes, K-Nearest Neighbors, Random Forest classifier, Logistic Regression, Bagging classifier, AdaBoost classifier, Gradient Boosting classifier, Stochastic Gradient Descent (SGD) classifier, Multi-layer Perceptron (MLP) classifier, XGBoost classifier, and Decision Tree classifier, as demonstrated. In this study, we employed the CHBMIT dataset of scalp EEG signals and tested our suggested methodology on the dataset's 22 participants. With superior performance and higher prediction accuracy, our suggested seizure prediction approach is able to reach 95.88% accuracy, 86.91% recall, 1% precision, and 1% sensitivity.