Under laboratory conditions, an experimental analysis is carried to evaluate the performance of thermal comfort-based controls compared to temperature-based controls. In this analysis, the Fanger model used to estimate a thermal comfort indicator that can be utilized to control a mechanical cooling system. Specifically, the Predicted Mean Value (PMV) of Fanger model was determined using two methods. In the first method, a commercially available thermal comfort sensor was used to measure an equivalent temperature which is then converted into PMV. In the second method, air temperature, air velocity, relative humidity, and mean radiant temperature were monitored and the PMV is calculated for a given occupant metabolic rate of occupant and clothing level. The results obtained for the second method was used to calibrate the thermal comfort sensor. In this paper, the results of a series of experiments are presented to determine if thermal comfort-based controls can save energy (while maintaining adequate thermal comfort) when compared to conventional control strategies based on maintaining indoor temperature within given set-points. It was found that when the PMV is set to 0.5, the thermal comfort based controls use 10% less energy than conventional controls. The energy savings are reduced to about 7% when the PMV is set to 0 (neural level).