Sleep is a vital physiological process crucial for overall health and well-being. Accurate identification and classification of sleep stages, including rapid eye movement (REM) and non-REM stages, are essential for understanding sleep disorders and diagnosing medical conditions. Traditional manual scoring by clinicians had limitations, leading to the development of automated methods using machine and deep learning techniques. This study explores automated sleep stage detection, mainly focusing on EEG signals. It reviews various algorithms, from traditional feature-based methods to state-of-the-art machine learning approaches. The research employs the MNE-Python library to preprocess and label EEG data from the Sleep-EDF dataset. We have mainly used a neural network model for the multi-class classification of sleep stages. Performance metrics like precision, recall, F1 score, accuracy, and Cohen’s Kappa are used to evaluate the models. The entire experiment was performed on thirty subjects. The results show promising classification accuracy with an accuracy of approximately 76.8% and a Cohen’s Kappa of 0.678, marking a significant step toward more efficient and scalable sleep stage detection. The study discusses potentially future directions, including multimodal data fusion and personalized modeling, to address challenges in sleep stage classification. Automated sleep stage detection improves our understanding of sleep patterns, disorders, and overall health.