Background The efficiency of the mandibular advancement device (MAD) in patients with obstructive sleep apnoea syndrome (OSAS) has been demonstrated. Nevertheless, the behaviour of the upper airway once MAD is placed and titrated, and its correlation with the apnoea‐hypopnoea index (AHI) is still under discussion. Objectives To analyse the morphological changes of the upper airway through a bi‐ and three‐dimensional study and correlate it with the polysomnographic variable, AHI. Methods Patients were recruited from two different hospitals for the treatment of OSAS with a custom‐made MAD. A cone‐beam computer tomography and a polysomnography were performed at baseline and once the MAD was titrated. Results A total of 41 patients completed the study. Treatment with MAD reduced the AHI from 22.5 ± 16.8 to 9.2 ± 11.6 (p ≤ .05). There was a significant increase of the total airway volume with MAD from 21.83 ± 7.05 cm3 to 24.19 ± 8.19 cm3, at the expense of the oropharynx. Moreover, the correlation between the improvement of the AHI and the augmentation of the volume of the upper airway was not statistically significant. Conclusions The oral device used in this prospective study increased the mean upper pharyngeal airway volume and significantly reduced the AHI. Future studies that measure the muscular tone are needed to completely understand the association between the AHI and the physiological and anatomical response of the upper airway.