Abstract: Objective To explore the risk factors for nocturnal hypoxemia in patients with obstructive sleep apnea (OSA) and construct an assessment model. Methods A total of 303 patients with OSA diagnosed by polysomnography (PSG) were selected as research objects. Based on PSG results, they were divided into case group (171 patients with hypoxemia) and control group (132 patients without hypoxemia). Materials were compared between the two groups; multivariable Logistic regression analysis was used to screen risk factors for nocturnal hypoxemia in OSA patients and construct a risk assessment model. Value of the model was evaluated by the area under the curve (AUC) of the receiver operating characteristic (ROC) curve and K-fold cross-validation. Results The case group had significantly higher values in age, body mass index (BMI), neck circumference, proportion of smokers, hemoglobin, red cell distribution width (RDW), apnea-hypopnea index (AHI), the longest apnea duration (LAD), snoring index, and percentage of apnea-hypopnea duration to total sleep time (AHT%) when compared to the control group (P < 0.05). Logistic regression analysis showed that high BMI, high RDW, long LAD, and high AHT% were risk factors for nocturnal hypoxemia in OSA patients (P < 0.05). AUC value (0.952) of the ROC curve of the assessment model constructed based on these factors was higher than the AUC values of BMI (0.833), RDW (0.780), LAD (0.866), and AHT% (0.898) alone. One hundred times of 10-fold cross-validation demonstrated that the assessment model had good generalization ability. Conclusion Nocturnal hypoxemia in OSA patients is associated with high BMI, high RDW, long LAD, and high AHT%. The model constructed based on these factors plays an important role in evaluating nocturnal hypoxemia in OSA patients.