To measure and track human performance and physiological state, a great number of biomedical data such as the electroencephalographic (EEG) signal is collected from the human body every day. For research and for medical diagnosis and treatment, understanding these signals is critical. There are many research and development projects Within this field, but there are weaknesses point, including their reliance on specific registration technologies, low accuracy and timing of implementations. In This paper a computeraided system are used for diagnosing epileptic seizures more accurately using electronic imaging data. The framework relies on identifying (EEG) signals for use in linear and non-linear applications. This study was conducted using six algorithms of machine learning algorithms which are logistic regression (LR), support vector machine (SVM), k-Nearest Neighbors (KNN), Gaussian Naive Bayes, Artificial Neural Networks (ANN), Principal Component Analysis(PCA). Also, improve surrogate data technique are used for feature extractions that depend on Fourier transforms (FT). It was concluded that the classifier (SVM) works better and showed high accuracy in the classification of data taken for all epilepsy cases.