The low pressure and oxygen in high-altitude areas bring severe challenges to the physiological and mental health of the plain people who enter the high-altitude area. At present, the evaluation of hypoxia in plateau populations mainly relies on subjective symptom description, and due to the fact that human hypoxia response is affected by multi-dimensional factors such as physical function, altitude of the source area, and residence time, conducting experiments with multiple working conditions leads to a large amount of workload, high costs, and subjective results. Therefore, the idea of replacing multi-dimensional complex adaptation factors with lump-sum “hypoxia adaptation level” was proposed in this study. Objective physiological parameters for hypoxia evaluation were obtained through experiments on various oxygen concentration conditions at the plateau. The Receiver Operating Characteristic Curve (ROC) of physiological indicators and oxygen sensation were established, area under the curve (AUC), asymptote significance, and correlation with oxygen sensation were analyzed, and evaluation indicators were screened. The K-means clustering method combined with Fisher discrimination was used to construct the oxygen adaptation level prediction model suitable for the alien population in the high-altitude area. Finally, the oxygen adaptation level of the population is divided into high oxygen adaptation level, moderate oxygen adaptation level, and low oxygen adaptation level, and the accuracy rate is as high as 94.96%. It can accurately determine the adaptation state of the alien population in the hypoxic environment, and it can provide a reference basis for the regulation of the indoor oxygen environment in the plateau.