Distribution network parameters are subject to change because of factors such as weather or load conditions. Then, because of factors such as limited installation of sensing devices and delayed data transmission, it is difficult to obtain accurate parameters of the distribution network, thereby affecting the accuracy of the traditional fault location method. Thus, establishing a digital twin model of the distribution network based on the model’s parameter self-correction technology, this paper puts forward a fault localization method of the distribution network in which the localization model is dynamically corrected with the parameter changes. Meanwhile, this paper builds a platform based on a digital twin server and RTDS, and realizes synchronous communication between a real-time physical model of a distribution network and a digital twin model. In the case of simulation, the distribution network fault location method based on digital twin technology is verified using the proposed digital twin platform. The results of the simulation show that the method can correct distribution network parameters in real time and significantly improve the speed and accuracy of distribution network fault location.