The invention introduces a feeder fault handling method, which comprises the following steps of: 1) conducting online search and locating a feeder fault; 2) conducting network topology analysis to draw a conclusion; 3) finding out the section of the fault; 4) searching for the boundary switch of the fault section and making a fault isolation plan; 5) defining non-fault power outage areas; 6) making power supply restoration plans for outage areas on the upstream side and the downstream side of the fault; 7) conducting safety evaluation to the downstream power supply restoration plan; and 8) executing the plans according to the safety evaluation obtained at the step 7. The feeder fault handling method has the advantages that: 1) the method is suitable for the automatic construction of feeders of a power distribution network which is wired in any way; 2) the practicability is high, the impedance parameters of the equipment of the distribution network are not required and the power flow calculation module is not required to be called; and 3) the power supply restoration can be separately conducted to the non-fault sections, so as to reduce the power outage range and decrease the poweroutage duration as much as possible.