Mandibular advancement device (MAD) is an alternative therapeutic option for CPAP to treat obstructive sleep apnea (OSA). While MAD showed the better adherence, patients with over moderate OSA have been treated more frequently with CPAP despite increasing positive evidence on the cardiovascular outcome with MAD, even in severe patients. Thus, more information is needed regarding the cardiovascular and symptomatic outcome of MAD treatment objectively compared to CPAP. Forty-five supine-dependent OSA patients (apnea–hypopnea index 20–40/h) were randomized to either CPAP or MAD and treated for 8 weeks and switched to another for 8 weeks. The primary endpoint was improvement in the endothelial function, indexed by the flow-mediated dilatation (FMD), and the secondary endpoint was the sleep-time blood pressure (BP). The duration of MAD use was evaluated objectively by an implanted adherence monitor. Treatment efficacy was also evaluated by home sleep monitor and a questionnaire about the symptoms. The adherence was not significantly different (CPAP vs. MAD: 274.5 ± 108.9 min/night vs. 314.8 ± 127.0 min/night, p = 0.095). FMD and sleep-time mean BP were not markedly changed from the baseline with either approach (CPAP vs. MAD: FMD, + 0.47% ± 3.1% vs. + 0.85% ± 2.6%, p = 0.64; BP, − 1.5 ± 5.7 mmHg vs. − 1.2 ± 7.5 mmHg, p = 0.48), although sleepiness, nocturia, and sleep-related parameters were similarly improved and more patients preferred MAD. As MAD and CPAP showed similar effects on cardiovascular outcome and symptomatic relief even with a comparable length of usage, we might expect MAD as an alternative treatment option for CPAP in this range of OSA group.