BackgroundThe outbreak of the highly infectious coronavirus disease 2019 (COVID-19) renders a huge physical and psychological risk to the public, especially to the medics. Additionally, self-leadership has proven to improve self-efficacy and mediate tension, such as nervousness and depression. Therefore, a cross-sectional survey was conducted to explore the association of self-leadership with acute stress responses (ASRs) and acute stress disorders (ASDs) in medics during the COVID-19 epidemic. MethodsSelf-reported online questionnaires were administered, and 627 participants were finally included. The data were analyzed using the univariate analysis and the logistical regression model to identify whether self-leadership and sociodemographic and epidemic characteristics were associated with mental health, including ASRs and ASDs. ResultsInitially, 790 medics responded. Of these, 627 remained after excluding for invalid questionnaires and those with a substantial amount of missing data. Therefore, the participation validity rate was 79.37%. Frontline medical staff (beta = 0.338; p < 0.001), possibility of infection among people around the medic being mild (beta = 0.141; p < 0.001), subjective estimation of epidemic duration being 3-6 months (beta = 0.074; p < 0.05), self-sets (beta = -0.022; p < 0.001), self-punishment (beta = 0.229; p < 0.001), belief hypothesis and evaluation (beta = -0.147; p < 0.05), and successful foresight (beta = 0.105; p < 0.05) were statistically significant with ASRs. Marital status [adjusted odds ratio (AOR) =1.813; 95% CI (1.141, 2.881); p = 0.012], being a frontline worker [AOR = 25.760; 95% CI (14.220, 46.667); p < 0.001], visiting Hubei in the previous 14 days [AOR = 3.656; 95% CI (1.500, 8.911); p = 0.004], self-punishment [AOR = 1.352; 95% CI (1.180, 1.548); p < 0.001], and self-dialogue [AOR = 1.256; 95% CI (11.063, 1.483); p = 0.007] were the risk factors for ASD. Conversely, having frontline medical staff in one's family [AOR = 0.523; 95% CI (0.297, 0.923); p = 0.025], self-sets [AOR = 0.814; 95% CI (0.715, 0.826); p = 0.002], and belief hypothesis and evaluation [AOR = 0.796; 95% CI (0.672, 0.943); p = 0.038] were the protective factors. ConclusionThe special working environment of the COVID-19 epidemic resulted in ASR and ASD. Notably, findings revealed a positive association between ASR symptoms and frontline medical staff, the subjective estimation of epidemic duration, self-punishment, and successful foresight. Nevertheless, marital status, having visited Hubei in the previous 14 days, and self-dialogue were the risk factors accounting for ASD symptoms. Surprisingly, having frontline medical staff in one's family, self-sets, and belief hypothesis and evaluation had potential benefits for ASD symptoms.