It is difficult for a distribution network to achieve accurate fault location because of its complex structure and many branches. Thus this paper proposes a distribution network fault location method based on subsystem division and injection current ratio. First, the network is divided into several independent subsystems according to the number of end nodes, and the equivalent virtual injection current of the head and end nodes is calculated according to any subsystem, and then the injection current of the fault point can be obtained. Then, assuming that the current is injected from different nodes of each subsystem, the fault subsystem is determined according to the error between the calculated value and the measured value of the end node voltage. Finally, in the fault subsystem, the relationship between the injection current ratio of the head and end nodes and the position of the real fault point is used to locate the fault. The proposed method only needs to configure a micro synchronous phasor measurement unit (μPMU) at the head and end nodes. This method has good economy. Through Matlab/Simulink simulation tests, the proposed method is not only less affected by fault type, transition resistance, fault position and fault initial phase angle, but is also insensitive to load variation and noise.