Aiming at the problems of complex fault mechanism and difficult to adapt to the change of topology and fault characteristics for the secondary equipment of smart substation failure, a fault location method of the secondary equipment in smart substation based on gated recurrent unit (GRU) is proposed. For the fault location objects such as merging unit, intelligent terminal, protection device, its sending and receiving network ports, and optical fibers, the alarm signals from related equipment are used when the secondary equipment fails to form an alarm signal set. Using the GRU network, each deep learning network fault location model for line, bus, and main transformer bays is established, and various training strategies for the secondary equipment fault location model are given. Through the case of a typical smart substation, and the effectiveness and accuracy of the proposed fault location method are verified by simulation experiments, and compared with the long short-term memory network and recurrent neural network, the proposed method can be more accurately and rapidly to locate secondary equipment in smart substation.