Rotor faults such as misalignment, shaft crack, rotor rub, bearing fault etc. lead to high vibration and are major cause of machinery shut down. Misalignment in rotating machinery shafts develops reaction forces and moments at the coupling. Conventionally, misalignment diagnosis is carried through vibration measurements. Especially, the existence of strong 2x component is widely accepted. However, there are other rotor-bearing faults which lead to significant 2x vibrations. Hence distinguishing misalignment is a challenging task using vibration signals alone. In view of these limitations by employing vibrations signals, present study focuses on torque measurements for misalignment diagnosis. Experimental results using torque sensor for different cases of misalignment at different frequency of operation are reported. Fourier and wavelet transforms are used to detect the misalignment fault. The variation of the fault characteristics with respect to time are reported for sampled signals using wavelet transforms. The study shows the possibilities of using torque measurements as useful and additional technique for the detection and monitoring of coupling misalignment. © 2014 Elsevier Ltd. All rights reserved.