Background We sought to describe the tendency and extent of high-sensitivity cardiac troponin I (hs-cTnI) changes in patients with fulminant myocarditis (FM) after admission and to explore the relationship between the in-hospital mortality of FM and the absolute and relative changes in hs-cTnI within 24 h and 48 h after admission. Methods In the retrospective study, the object are patients diagnosed with FM in our single centre. The value of cardiac troponin was recorded after patients admitted to hospital in succession. The absolute and relative changes in hs-cTnI within 24 h and 48 h were described as range distributions. Receiver operating characteristic (ROC) curve and Cox analyses were performed to determine the relationship between in-hospital mortality of FM and hs-cTnI changes. Results A total of 83 FM patients admitted to our centre from January 1, 2010 to December 31, 2019 were included; 69 patients survived and 14 patients died. In the survival group, 78% of patients experienced a decline in hs-cTnI within 24 h, while 36% of the mortality group exhibited a declining tendency in hs-cTnI (P = 0.003). Nearly 60% of survival group had a 0-2000 ng/l reduction in troponin from baseline within 24 h of admission. However, troponin levels of 50% of patients in the mortality group were 0-10,000 ng/ L higher than baseline 24 h after admission. Multivariable logistic analysis revealed that the declining tendency of hs-cTnI within 24 h, in addition to time from onset to admittance to hospital, intravenous immunoglobulin treatment and the abnormal level of creatinine, were associated with the in-hospital mortality of FM (for the declining tendency of hs-cTnI within 24 h, OR = 0.10, 95% CI 0.02-0.68, P = 0.018). The ROC curve revealed optimal cut-off values of - 618 ng/l for absolute change within 24 h (AUC = 0.800, P < 0.01), - 4389 ng/l for absolute change within 48 h (area under the curve = 0.711, P < 0.01), - 28.46% for relative change within 24 h (AUC = 0.810, P < 0.01), and - 52.23% for relative change within 48 h (AUC = 0.795, P < 0.01). Absolute changes and relative changes in hs-cTnI within 24 h and 48 h were strong predictors of in-hospital mortality by Cox regression analysis after adjustment for sex, time from onset to admission, and occurrence of ventricular tachycardia or ventricular fibrillation. Conclusion Most FM patients who survived experienced a decline in hs-cTnI within 24 h. The absolute and relative changes in hs-cTnI within 24 h and 48 h were strong predictors of in-hospital mortality.