We investigated myocardial performance concerning obstructive sleep apnea (OSA) severity and the benefits of continuous positive airway pressure (CPAP) therapy. In this randomized sham-controlled trial, 52 patients (mean age, 49 years; 92%, males; mean AHI, 59) with severe OSA were randomly assigned to receive either CPAP or sham treatment for 3 months. The severity of OSA was determined using the apnea/hypopnea index (AHI), oxygen desaturation index (ODI), percentage of sleep time below 90% oxygen saturation (T90), and average O saturation during sleep (mean SpO). We compared the changes in myocardial work after 3 months of CPAP (n = 26) versus the sham group (n = 26) at rest and during an exercise stress test. Unlike AHI or ODI, indices of hypoxemia including T90 and mean SpO were significantly correlated with global constructive work, as defined by work of left ventricle (LV) that contributes to LV ejection during systole (T90, β = 0.393, p = 0.012; mean SpO, β = 0.331, p = 0.048), and global wasted work (GWW), as defined by work of LV that does not contribute to LV ejection (T90, β = 0.363, p = 0.015; mean SpO, β =  - 0.370, p = 0.019). After 3 months, GWW decreased (80.0 ± 49.2 to 60.8 ± 26.3, p = 0.009) and global work efficiency increased (94.0 ± 4.5 to 95.7 ± 2.0, p = 0.008) in the CPAP group compared to those in the sham group. At the 3-month follow-up exercise stress echocardiography, worsening of GWW during exercise was significantly decreased in the CPAP group compared to that in the sham group (p = 0.045 at 50 W). Hypoxemia indices were closely associated with myocardial performance in patients with severe OSA. CPAP treatment for 3 months improved left ventricular myocardial performance by decreasing wasted work and increasing work efficacy compared to the sham treatment.