Objective:To investigate the clinical value of eosinophil count in predicting the progression and prognosis of COVID-19.Methods:In this retrospective cohort study, 115 patients confirmed COVID-19 were enrolled in Taizhou Public Health Medical Center, Taizhou Hospital, Zhejiang Province, China, from January 22, 2020 to February 12, 2020. The subjects were divided into non-severe ( n=90) and severe ( n=25) groups, of which medium age was 46 years old, including 65 male and 50 female subjects. The value of eosinophil count in reflecting the severity of COVID-19 was tested with a receiver operating characteristic curve (ROC) analysis; Correlation analysis of eosinophil count at the admission with length of stay (LOS) were studied. GraphPad Prism6 and SPSS 19.0 statistical software were applied to data analysis. Differences among groups were analyzed with the Mann-Whitney U test. Results admission, Eosinophil counts of severe patients were significantly lower than those of non-severe ( P<0.001). The eosinophil count remained below normal for 1-7 days after admission and rose to normal by 21 to 28 days; The area under the ROC curve (AUC) of eosinophil to COVID-19 severity was 0.781(95% CI:0.693-0.869); The Eosinophil count at admission was negatively correlated with the patient's length of stay. Conclusions:The decrease of Eosinophil count can serve as a risk factor for auxiliary diagnosis in severe COVID-19 patients. The dynamic monitoring of eosinophils is useful for prognostic purposes.