Abstract: Snoring is the most common symptom in the patients with obstructive sleep apnea hypopnea syndrome (OSAHS). It provides important diagnostic information for patients. In this study, acoustic features such as Mel-frequency cepstral coefficients (MFCC), linear predictive coefficients (LPC), and spectral entropy (SE) are extracted from the snoring sounds of patients throughout the night. The correlation between snoring characteristics and the apnea hypopnea index (AHI) is evaluated through correlation analysis. An AHI prediction model based on gradient boosting regression is investigated, and its performance is compared with that of other models. The results indicate a strong correlation between some characteristics of SE and MFCC with AHI, which have a correlation coefficient greater than 0.6. Compared to the linear model, the gradient boosting regression model demonstrates better prediction performance with a correlation coefficient of 0.813 between predicted AHI and measured values. These findings manifest that it is feasible to predict AHI based on the acoustic characteristics of snoring, providing valuable reference for diagnosing OSAHS patients.