This paper presents detection of broken rotor bar fault using Stockwell transform of the stator current signals. The transform gives complex two-dimensional information pertaining to time and frequency. Entropy of the matrix with respect to time domain has been proposed as the fault detection parameter. Entropy is found to be useful for the detection as well as for the estimation of the severity of broken rotor bar faults. The results have been compared with the energy of the matrix. The results show that energy can only detect the fault unlike entropy. Thus, entropy can be used as a fault diagnostic parameter. The results have been validated on the experimentally recorded current signals for five different load conditions; from no-load to full-load. © 2020 IEEE.