The question of why we age and finally die has been a central subject in the life, medical, and health sciences. Many aging theories have proposed biomarkers that are related to aging. However, they do not have sufficient power to predict the aging process and longevity. We here propose a new biomarker of human aging based on the mass-specific basal metabolic rate (msBMR). It is well known by the Harris-Benedict equation that the msBMR declines with age but varies among individual persons. We tried to renormalize the msBMR by primarily incorporating the body mass index into this equation. The renormalized msBMR (RmsBMR) which was derived in one cohort of American men (n = 25,425) was identified as one of the best biomarkers of aging, because it could well reproduce the observed respective American, Italian, and Japanese data on the mortality rate and survival curve. A recently observed plateau of the mortality rate in centenarians corresponded to the lowest value (threshold) of the RmsBMR, which stands for the final stage of human life. A universal decline of the RmsBMR with age was associated with the mitochondrial number decay, which was caused by a slight fluctuation of the dynamic fusion/fission system. This decay form was observed by the measurement in mice. Finally, the present approach explained the reason why the BMR in mammals is regulated by the empirical algometric scaling law.