Objective: To analyze the value of continuous finger pulse oxygen saturation (SpO2) monitoring in guiding oxygen therapy in patients with pneumonia complicated with hypoxemia. Methods: 101 subjects with viral pneumonia complicated with hypoxemia were enrolled, 56 with continuous finger pulse oximetry monitoring and 45 with routine monitoring. Two groups were compared about hospital days, arterial blood gas counts, course of oxygen therapy. Effects of two monitoring on blood oxygenation and inflammatory factors [interleukin 6（IL-6）and C-reactive protein （CRP）] on day（7±1） were evaluated an ANCOVA model. Data from the continuous monitoring group was divided into escalation and no escalation groups for subgroup statistical analyses according to whether an oxygen therapy escalation event occurs within 48h of admission. Correlations between monitoring parameters and inflammatory and immune factors were analyzed. The monitoring data of patients was analyzed who had been upgraded to bilevel positive airway pressure（BiPAP） oxygen therapy who had another oxygen therapy escalation event under 48h. The diagnostic value of pulse oximetry monitoring parameters for the diagnosis of early oxygen therapy escalation events was explored using receiver operating characteristic（ROC） analysis. Results: The ABGs counts, the length of oxygen therapy initiation escalation and reaching the maximum parameters were smaller in the continuous monitoring group than routine group（P＜0.05）. Continuous monitoring versus routine monitoring was associated with significantly greater inflammatory factors improvements in oxygenation index on day7±1 (LS mean difference, 29.52(95% CI:2.27 - 56.76， P=0.034)), IL-6(-4.83(95% CI:-9.58 to -0.83， P=0.046)) and CRP(-4.83(95% CI:-9.58 to -0.83， P=0.046)).The differences in monitoring parameters in the continuous pulse oximetry monitoring group in lowest SpO2（SpO2L）, mean SpO2（mSpO2）, the percentage of time with SpO2 below 86%、88%、90%、92%、and 94%（T86、T88、T90、T92、T94）, standard deviation were statistically significant between the upgraded group and the no-upgraded group（P＜0.05）. IL-6, CRP and immunoglobulin G were significantly correlated with pulse oximetry parameters. In addition, continuous monitoring parameters on BiPAP days showed events of sudden oxygen desaturation accompanied by increasing pulse rate. Such hypoxic events were found to be associated with BiPAP offline, which can be used to diagnose escalation events. Conclusion: Continuous finger pulse oximetry monitoring parameters T92 can identify the risk of progressive exacerbation of hypoxic events at an early stage. The area of decline of oxygen desaturation under BiPAP day offline can be used as a predictor of oxygen therapy escalation events. Continuous finger pulse oximetry monitoring can help guide decisions about oxygen therapy regimens in patients with pneumonia complicated by hypoxemia.