Eicosapentaenoic acid (EPA) has antiarrhythmic effects. The J-wave on an electrocardiogram is associated with a high incidence of ventricular tachycardia/fibrillation (VT/VF). We evaluated relationships between EPA and J-waves, and their involvement in the occurrence of VT/VF in acute myocardial infarction (AMI). Two hundred consecutive patients undergoing successful percutaneous coronary intervention within 12 h after AMI onset were enrolled. Serum EPA level and J-waves at admission were evaluated. The patients were divided into two groups according to the optimal cutoff value (2.94) of serum EPA level (% of total fatty acids): LOW (<2.94, 61 ± 11 years, n = 103) and HIGH groups (≥2.94, 70 ± 13 years, n = 81). J-waves were observed more frequently in the LOW (36/103, 35 %) than in HIGH group (16/81, 20 %) (P = 0.020). The 30-day incidence of VT/VF including those occurring before admission was higher in the LOW (19.5 %) than in HIGH group (6.2 %) (P = 0.009). The patients with J-waves showed a higher incidence of VT/VF (23.1 %) than those without J-waves (9.9 %) (P = 0.019). Kaplan-Meier analysis showed that the highest incidence of VT/VF was observed in the LOW with J-wave group (27.8 %), followed by the LOW without J-wave (15.0 %), HIGH with J-wave (12.5 %), and HIGH without J-wave (4.6 %) (P = 0.013). Cox proportional hazard analysis revealed that Killip grade and low serum EPA level or presence of J-waves were significantly associated with the incidence of VT/VF. Low serum EPA level is a risk for incidence of VT/VF in the acute phase of myocardial infarction. Involvement of the J-wave and its possible link with EPA in the pathogenesis of ischemia-induced VT/VF are suggested.