Sleeping disorders may lead to severe condition in the human’s life when it is not treated properly. Various sleeping disorders are addressed here and we are trying to predict at the earliest using machine learning algorithm to save the diseased. The raw EEG signals are collected from the Physionet for the sleep disorder patients and the signals underwent preprocessing. The transformed and decomposed signals are used to extract the valuable characteristics of the signal. Several features are extracted for both normal and abnormal signals and they feed for classifier for the prediction process. Here, some of the neural network frame work are involved in the prediction process. The usefulness of the artificial neural network is emphasized in this study. Totally, 16 number of sleeping disorder patients EEG (Electro-Encephalo Gram) data are processed and the features are selected for the training the network and testing is also done for the abnormality prediction. There are 30 number of normal sleep data that are used for this research. The training and testing phase of the study followed the 80–20 scheme for validation of the machine learning algorithm. The performance of the classifier in terms of sensitivity, specificity, accuracy, and false positive value has also been arrived at. The use of EEG-based sleep disorder detection has important economic benefits, making diagnosis more affordable and accessible.