The three-phase induction motor is known as one of the most widely-used machine in the manufacturing industry. Electrical and mechanical fault of this machine is able to cause breaking down the plant facilities that leads significant productivity losses. In this paper, we will propose the fault detection method of induction motor by using deep neural network with measurement data of the electrical current signals to provide supervised classification of 2 types ofmotor fault, rotor broken and stator conductor fault. We also studied the size of data-set in faulty state to train the model for fault detection.