Abstract: Objective To understand the sub-health status of undergraduate nursing students and predict their characteristic population by the classification and regression tree model. Methods From December 2021 to March 2022,363 nursing students enrolled in a university inXi'an City were collected by the convenience sampling,the sub-health status of nursing students were investigated by the suboptimal health status questionnaire-25（SHSQ-25）,and the data were analyzed. Results The sub-health detection rate of nursing students was 90.6%. There were statistically significant differences in the sub-health scores among undergraduate nursing students with different place of birth,self-rated stress level,and hours of sleep per night（all P<0.05）. The highest sub-health score[（80.57±16.12）points] was found in undergraduate nursing students with insomnia and staying up late habit,mental state was seriously or extremely affected by staying up late,and staying up late more than three times in a week,and the the lowest sub-health score[（46.18±10.94）points] was found in undergraduate nursing students who did not stay up late in the recent week. Conclusion The sub-health status of undergraduate nursing studentsin a university in Xi'an City is moderate or above,and affected by the grade,stress level,and the habit of staying up late,so the nursing educators should conduct scientific and effective health-related activities and knowledge lectures to improve the health awareness and status of undergraduate nursing students according to theirplace of birth,insomnia status,and reasons for staying up late.

Key words: Nursing students, Sub-health, Classification and regression tree, Prediction