Sleep disorders have a significant impact on human health and early diagnosis is crucial for the improvement of quality of life. In this work, machine learning models that classify sleep disorders have been developed from the dataset of 374 participants that comprises features relating to sleep health and lifestyle, such as sleep duration, physical activity, stress levels, and BMI. It has used many machine learning algorithms, like GNB, K-NN, SVM, RF, Logistic Regression, and ANN/MLP classifier. Dataset was highly preprocessed to deal with missing values and normalize features, thereby offering the best performance to the models. Feature engineering and optimization through a genetic algorithm improved the predictive ability of these models. GA was highly effective especially in the extraction of the most relevant features, improving the classification accuracy, and dealing with the problem of small dataset size.The best accuracy by GNB was achieved in combination with GA, namely, 94%, compared to all other models: SVM 93%, Logistic Regression 92%, and ANN 93%. Importantly, issues like dealing with missing data and feature extraction, where meaningful features such as systolic and diastolic blood pressure levels had to be extracted, were addressed in order to enhance the outcome. Our results show that the proposed model effectively detects patterns in the database, hence providing consequent results that can be used in real time by healthcare professionals to diagnose sleep-related conditions. This consequently decreases manual effort toward diagnosis and improves its timing. It also allows for scalability, which may fit easily into wearable health devices and clinical decision systems that are intended to prevent long-term health risks due to sleep disorders left untreated. The proposed model facilitates automated classification of sleep disorders. This will be a valuable tool for advancing healthcare outcomes and fostering future innovations in sleep medicine.