Improving the fault location accuracy of high voltage transmission line is helpful to find the fault location and repair it in time, and is of great significance to improve the power supply reliability, and guarantee a safe and stable operation of power grid. In this paper, the BP neural network is adopt to comprehensively integrate the fault location information of different ranging method, and then evaluate the fault distance optimally to improve the accuracy of fault location. Two integrated models that take fault distance and fault location error as the output respectively, are built, and the simulation calculation and the actual ranging data are utilized as training samples to train the neural network, and the applicability of these two models are analyzed. The results of practical application show that the proposed fault location comprehensive optimization method can make full use of the advantage of various ranging method, and can effectively improve the accuracy of fault location.