This is the first ever proposal to use skin hardness as a physiological sign by which to estimate human thermal status and to verify its effectiveness and independence in relation to the two conventional signs: skin temperature and skin conductance. We propose a novel TSV model adding skin hardness to the conventional TSV model for better estimation of human thermal status with higher accuracy and lower error. We survey individual thermal sensation from 30 subjects under four different thermal conditions (normal, warm, hot, and cold); while measuring skin hardness along with the two conventional physiological signs. The novel model for estimation of thermal status from all three signs increases R2 by 17.4% and decreases error by 23.5%, compared to the conventional model using two signs. The novel TSV model has potential for applications to human-machine interaction systems for better estimation of human thermal status.