obstructive sleep apnea (OSA) is a prevalent sleep related breathing disorder associated with several anatomical abnormalities of the upper airway. acoustic parameters of human speech are influenced by properties of the vocal tract, which includes the upper airway. We hypothesize that it is possible to differentiate OSA patients from non-OSA (healthy) subjects by analyzing potential patients' speech signals. using speaker recognition and signal processing techniques, we designed a system for classifying a given speech signal into one of the two groups. the database for this research was constructed from 92 subjects who were recorded reading a one-minute speech protocol immediately prior to a full polysomnography study; one hundred and three acoustic features were extracted from each signal; seven independent Gaussian mixture models (GMM)-based classifiers were implemented; a fusion process was designed to combine the scores of these classifiers and a validation procedure took place in order to examine the system's performance. specificity and sensitivity of 91.66% and 91.66% were achieved for the male population; and 88.89% and 85.71% were achieved for female population, respectively. such a system can be used as a tool for initial screening of potential OSA patients.