Sleep Apnea is a clinical condition related to problematic breathing during the course of sleep. Nocturnal polysomnography (NPSG) is the traditional method to detect sleep apnea but it is costly and inconvenient. Research directions are turning towards signal processing based sleep apnea detection. This study focuses on a simple, fast and memory efficient method for detection of sleep apnea using Electrocardiogram (ECG) data from MIT online database. Discrete wavelet packet transform (DWPT) has been employed on RRI intervals (RRI), in band of interest (FOI) to highlight the frequency pattern differences in normal and abnormal cases. Cases based classification is 100 accurate. © 2011 IEEE.