Schizophrenia is a psychiatric disorder that negatively affects both mental function and behavior. Identifying it in its early stages allows for more effective treatment, which contributes to improving patients’ quality of life. In recent decades, numerous studies have attempted to clarify the neurobiological mechanisms that give rise to its clinical manifestations, using tools such as electroencephalography (EEG). This technique, in addition to being non-invasive, is accessible for recording a person’s brain activity. The present research proposes to use a public dataset containing 1144 spectral power density (PSD) and functional connectivity (FC) features for the delta, theta, alpha, beta, high beta, and gamma frequency bands from EEG readings of 212 patients with and without schizophrenia. This dataset was preprocessed, and the Boruta feature selector was applied, which selected 23 of the total features. Holdout validation and cross-validation tests were performed, and Logistic Regression, Random Forest, AdaBoost, and Gradient Boosting were implemented as classification algorithms. The best-performing model was Logistic Regression, with an accuracy of 0.9219 and an area under the curve (AUC) of 0.9717. The results of this study indicate that machine learning algorithms applied to EEG signals have the potential to identify severe psychiatric disorders, such as schizophrenia. The Boruta algorithm was able to minimize the amount of characteristics by over 97%.