An estimation of myocardial mechano-energetic efficiency (MEE) per unit of left ventricular (LV) mass (MEEi) can significantly predict composite cardiovascular (CV) events in treated hypertensive patients with normal ejection fraction (EF), after adjustment for LV hypertrophy (LVH). We have tested whether MEEi predicts incident heart failure (HF), after adjustment for LVH, in the population-based cohort of a "Strong Heart Study" (SHS) with normal EF. We included 1,912 SHS participants (age 59 ± 8 years; 64% women) with preserved EF (≥50%) and without prevalent CV disease. MEE was estimated as the ratio of stroke work to the "double product" of heart rate times systolic blood pressure. MEEi was calculated as MEE/LV mass, and analyzed in quartiles. During a follow-up study of 9.2 ± 2.3 years, 126 participants developed HF (7%). HF was preceded by acute myocardial infarction (AMI) in 94 participants. A Kaplan-Meier plot, in quartiles of MEEi, demonstrated significant differences, substantially due to the deviation of the lowest quartile (p < 0.0001). Using AMI as a competing risk event, sequential models of Cox regression for incident HF (including significant confounders), demonstrated that low MEEi predicted incident HF not due to AMI (p = 0.026), after adjustment for significant effect of age, LVH, prolonged LV relaxation, diabetes, and smoking habits with negligible effects for sex, hypertension, antihypertensive therapy, obesity, and hyperlipemia. Low LV mechano-energetic efficiency per unit of LVM, is a predictor of incident, non-AMI related, HF in subjects with initially normal EF.