Mental stress assessments have been performed based on evaluation of autonomic nerve functions or physiological active substances. However, they have each shortcoming: the former is not always reliable due to complex responses of the autonomic nerves to various physical activities without emotional factors, and the latter needs biochemical assays which is an impediment to daily use. The present study provides a method using heart-rate variability in the very-low frequency range of below 0.04Hz (VLF) and demonstrates that VLF can be used as an effective stress marker throughout various field trials with continuous vital and behavioral records. It was also found that VLF changes responding to strong stress imposition while other spectral powers of the heart-rate variability at above the upper threshold of 0.04Hz didn’t show any correlates. This finding potentially makes the presented method much more reliable.