Scientists concentrate on the assessment of the micro and macro structure of sleep and the associated physiological activities in sleep. Their achievements heavily rely on the use of technology. Utilising the conventional method known as manual sleep stage scoring, is tedious and time-consuming. Thus, there is a significant need to create or develop a new automatic sleep stage detection system to assist the sleep physician in evaluating the sleep stages of healthy or non-healthy subjects. The main aim of this pilot study is to develop an algorithm for automatic sleep stage detection based on electrooculography (EOG) signals. 10 patients with periodic limb movements of sleep (PLMS), 10 patients with sleep apnea hypopnea syndrome (SAHS), and 10 healthy control subjects were utilised in this study. Numerous features were extracted from EOG signals such as cross-correlation, energy entropy, Shannon entropy and maximal amplitude value. K-Nearest Neighbor was used for the classification of sleep stages. An overall agreement between visual and automatic detection of sleep stage was estimated by 80.5% with Cohen’s Kappa 0.73. As a result, electrooculography (EOG) signals applied in the automatic sleep stage detection has shown a significant advantage. Knowing that fewer channels can be used to accurately detect sleep stages, it can be applied in ambulatory sleep stage recording and detection.