A silent syndrome known as Sleep Apnea Syndrome (SAS) affects between 2% and 4% of the population. SAS is a breathing disorder in which the patient does not breathe for more than 10 s, causing changes in heart rate and blood oxygenation. But the most worrying are the derived diseases that develop silently, such as hypertension, diabetes, cardiac arrhythmia, imbalance in the endocrine system, and neurological problems, among others. An accurate diagnosis is necessary to begin treatment for SAS, and the gold standard test is Polysomnography (PSG), a costly and time-consuming exam. In this context, this project aims to create a SAS screening methodology with accuracy, sensitivity, and low cost based on processing the QRS Complex (QRS) of a single Electrocardiogram (ECG) channel acquired from any device, regardless of the sampling frequency. The operation flow consisted of separating the data into sets with apnea and without apnea, applying a Modified Pan-Tompkins algorithm to detect the R peaks and disregarding the rest of the signal, extracting statistical features and coefficients derived from the Discrete Wavelet Transform (DWT) based on Heart Rate Variability (HRV) and apply the features in three classifiers: Residual Network (ResNet), Model based on Decision Variable (DV) and Support Vector Machine (SVM) with a 10 kfold cross validation. The results of the ResNet and DV-based classifiers presented low accuracy and low sensitivity, with high variance between validations, unlike the SVM classifier, which presented an accuracy of 92.11% and sensitivity of 95.99%, a result that presents security in a screening. The proposed method differs from those found in the literature, as it is independent of the signal sampling frequency and can be applied directly to any equipment regardless of the resolution of the signal obtained.