The invention provides a distribution network fault artificial intelligence analysis method. The method comprises the steps of S1, opening a distribution network automation system, and acquiring corresponding fault information; S2, carrying out visual representation on an GIS edge graph according to the tripping information, so as to obtain a tree barrier line segment which is preferentially checked; S3, analyzing tour inspection photo or a photo shot by an unmanned aerial vehicle through software, to obtain the type of the tree, analyzing the moment when the tree grows into a safety distanceof equipment according to the growth speed of the tree species, and warning the operation and maintenance personnel of the distribution network; S4, obtaining equipment and line segments which are preferentially checked in combination with defect recording and hidden danger recording of the production system; S5, obtaining distribution transforms preferentially checked according to heavy overloaddata of automatic distribution transforms of a distribution network; S6, obtaining lightning stroke equipment and line segments which are preferentially checked in combination with a lightning positioning system, and S7, displaying the lightning stroke equipment and line segments which are preferentially checked to the operation and maintenance personnel of the distribution network and providing atroubleshooting priority list.