The unique torque measurement method presented in this paper provides a useful means to assess the power output of a gas turbine and/or a steam turbine in power generation operations. It measures the torque with a high degree of accuracy on a rotating load coupling shaft connecting a driving turbine to a power generator during power plant operations. The measured torque can be used to verify various performance modeling tools based on a heat balance approach. Performance validation aids in reducing conservative benefit margins put onto upgrade packages due to less uncertainty in the simple cycle and combined cycle models. This technique is especially valuable in a combined cycle operation in which the power outputs of a gas turbine and a steam turbine coupled to the same shaft need to be determined separately. Unlike the conventional strain gage technique, this optical method offers a non-contact approach that does not require any embedded sensors or lead wires on the rotating shaft; it rather requires a few discrete targets installed on the shaft, a rigid frame to hold the optical probes steady for signal capturing, and a rotor reference or once per revolution sensor. Torque measurement is achieved by measuring the time arrival of the targets on the rotating shaft and then converting timing data into the angle of twist, which can then be converted to torque using predetermined torsional stiffness obtained from a torque calibration test on the same shaft. This method has been demonstrated in a full-scale gas turbine test, from which the results indicate a good correlation between torque measurement and performance modeling. Additionally, comparative tests involving two independent measurements on the same rig have been carried out in the laboratory and the results provide further evidence that confirms the accuracy and stability of this method.