Objective To investigate the therapeutic effect of chimeric antigen receptor (CAR) T cells on myocardial infarction (MI) and observe the impact of different administration routes on CAR-T cell cardiac toxicity. Methods Twenty-four SCID Beige immunodeficient mice, which successfully established a myocardial infarction model, were randomly divided into Hank’s balanced salt solution (HBSS) group, CAR-TIM group, CAR-TIV group, and CAR-TIMIV group (n＝6 for each group). In addition, a Sham group (n＝6) was set up. The Sham group, HBSS group, CAR-TIM group, and CAR-TIMIV group were injected with HBSS or CAR-T cells via intramyocardial injection on the 7th day after modeling. The Sham group, HBSS group, CAR-TIV group, and CAR-TIMIV group were injected with HBSS or CAR-T cells via tail vein injection on the 7th day and 14th day after modeling. After 28 days of modeling, the electrical physiological indicators of the mice were observed, and the myocardial tissues were subjected to Masson staining to calculate the area of myocardial infarction. Results Compared with the Sham group, the HBSS group had significant increases in myocardial infarction size and incidence of arrhythmia (P＜0.05). Compared with HBSS group, different routes of CAR-T cell therapy significantly reduced the area of myocardial infarction (all P＜0.05) and significantly increased the incidence of arrhythmia (all P＜0.05). There was no significant difference in the effect of CAR-TIV and CAR-TIM groups on the area of myocardial infarction and the incidence of arrhythmia. Compared with the CAR-TIV group,the area of myocardial infarction significantly reduced in the CAR-TIMIV group (P＜0.05), with no significant difference in the incidence of arrhythmia between the two groups. Conclusions Intravenous and local myocardial injection of CAR-T cells can effectively reduce the myocardial infarction area but increase the incidence of arrhythmia. The mechanism needs further study.