Snoring is a typical symptom of obstructive sleep apnea-hypopnea syndrome (OSAHS). In this study, an effective OSAHS patient detection system based on subject independence using snoring sounds is presented. A series of acoustic features including Mel cepstral coefficient, bark sub-band feature, perceptual linear prediction coefficient, pitch frequency, formant, 800Hz power ratio, spectral entropy, and gammatone cepstral coefficient are extracted. The feature selection based on the Fisher ratio is conducted and the Gaussian mixture models (GMM) of OSAHS patients and simple snorers were established respectively. For the test subject, the difference in log-likelihood probabilities of the two GMMs is calculated to identify the OSAHS patients. The proposed model achieves 90% accuracy and 95.65% precision based on leave-one-subject-out cross-validation using selected features with a dimension of 100. The average prediction time of the proposed model is 0.134 ± 0.005s. The promising results demonstrate the effectiveness and low computational cost of diagnosing OSAHS patients using snoring sounds at home.