: Temperature profile is one of the most important factors which can be used to predict remaining lifetime and fault progress of electrical machines. Machine lifetime depends on insulation lifetime whichbasicallydepends on temperature. Torque and power density of machine depends on current density, too. Losses and temperature are limitting factors of current density in electrical machines. Thermal analysis is a tool for machine cooling design, optimization and also can be used forfault diagnosis in electrical machines. One of the most frequent faults that occur in an electric machine is stator winding fault. Thiskind of fault is made by stator windings’ insulation breakdown due to over voltage, temperature stress, etc. Induced voltage in shorted tur ns causeslarge circulating current which makes high loss and heatin fault region. So, temperature profile on motor frame can be used as a measure of fault diagnosis of stator winding faults. In this thesis, in order to obtain temperature profile on motor frame, a coupled electromagnetic-thermal model is used.This profile iscomputed for differentstates such as healthy motor and motor with electrical faults (inter-turn short circuit fault, phase-to-phase and phase-to-ground fault)and compared with healthy motor. By comparison of simulation results,it is concluded that temperature profile for each fault is different from other types. So, with analysis of obtained profiles,fault type can be identified. Artificial neural network is used for detection of faults by extracted features from temperature profiles. Analysis of the results shows that standard deviation statistical parameter increases as fault severity rises. So, this parameter is used as an index for determination of fault severity. Keywords: thermal analysis, fault diagnosis, stator winding fault,temperature profile, artificial neural network