Visual fatigue syndrome is a set of visual and ocular disturbances resulting from prolonged use of digital screens. It is estimated that 70% of the global population suffers from this syndrome. In this study, an electrooculography (EOG) device was developed with electrodes integrated into an ophthalmic frame. The methodology included EOG signal acquisition, filtering, feature extraction, and data augmentation, followed by evaluation in three machine learning models: Artificial Neural Networks (ANN), Support Vector Machine (SVM), and K-Nearest Neighbors (KNN). The results of accuracy obtained were 92.04%, 75%, and 62.50%, respectively, demonstrating the feasibility of the approach for detecting asthenopia. These findings represent an advance in the early detection of visual fatigue through EOG and machine learning techniques, especially highlighting the effectiveness of ANNs in this field. It is suggested to expand the dataset and diversify subjects for more robust validation of the classification models and to explore broader clinical applications.