Sleep apnea syndrome (SAS) is a disorder in which breathing events such as apnea or hypopnea during sleep causes sleepiness and fatigue. SAS is a risk factor for coronary artery disease ( angina and myocardial infarction), atrial fibrillation, stroke, etc. The prevalence of SAS is reported to be 2-7% in adults; however, more patients with less subjective symptoms is estimated to have SAS. SAS is generally diagnosed by polysomnography (PSG) in specialized sleep institutes. However, PSG is performed only in a limited number of centers. Therefore, a screening method for SAS is needed to be developed. We focused on heart rate variability related to respiratory events and developed a screening method for SAS using neural networks. We examined a large PSG data set (N = 938) and attempted to detect SAS using long-term and short-term memory for heart rate data. Severe SAS was detected with an area under the curve (AUC) of 0.92, a sensitivity of 0.80, and a specificity of 0.84. We aim to develop a convenient screening method using a wearable device for the early detection of SAS.