Hypoxemia and obstructive sleep apnea (OSA) have been recognized as a threat to life. Nonetheless, information regarding the association between pre-dialytic pulse oximeter saturation (SpO2) level, OSA and mortality risks remains mysterious in patients with maintenance hemodialysis (MHD). Bioclinical characteristics and laboratory features were recorded at baseline. Pre-dialytic SpO2 was detected using a novel microchip LED oximetry, and the Epworth Sleepiness Scale (ESS) score greater than 10 indicated OSA. Non-adjusted and adjusted hazard ratios (aHRs) of all-cause and cardiovascular (CV) mortality were analyzed for pre-dialytic SpO2, OSA and potential risk factors. During 2152.8 patient-months of follow-up, SpO2 was associated with incremental risks of all-cause and CV death (HR: 0.90 (95% CI: 0.82–0.98) and 0.88 (95% CI: 0.80–0.98), respectively). The association between OSA and CV mortality was significant (HR: 3.19 (95% CI: 1.19–9.38). In the multivariate regression analysis, pre-dialytic SpO2 still had an increase in all-cause and CV death risk (HR: 0.88 (95% CI: 0.79–0.98), 0.82 (95% CI: 0.71–0.96), respectively). Considering the high prevalence of silent hypoxia in the post COVID-19 era, a lower pre-dialytic SpO2 level and severe OSA warn clinicians to assess potential CV risks. In light of clinical accessibility, the microchip LED oximetry could be developed as a wearable device within smartphone technologies and used as a routine screen tool for patient safety in the medical system.