The invention discloses a power distribution network fault intelligent searching and positioning method based on multi-source data fusion, and relates to the field of the power distribution network fault diagnosis. The method comprises the following steps: determining a fault occurrence high-probability region, establishing a fault data matrix related to the fault occurrence high-probability region and a network topological node according to the multi-source data of the power distribution network for fault region positioning; constructing a fault description matrix by utilizing the uploaded information of monitoring terminals of a circuit breaker, a section switch and an interconnection switch at the fault moment by the fault data matrix, then performing fault region layer positioning; performing line layer positioning in the determined fault section range, and calculating by utilizing electric quantity information of the nodes at two ends of the fault section, thereby obtaining the precise fault location. Under the support of a judgment result of the region layer, the precise positioning of the fault can be quickly realized, the adopted fault positioning principle is simple and reliable, and the strong anti-transition resistance capacity is provided; a sampling frequency requirement is low, and the fault positioning precision can be improved by improving the sampling frequency.