In order to improve the speed and fault tolerance of feeder fault location in a distribution network, a fault location method based on an integer linear programming model is proposed. The missing and false information of each Feeder Terminal Unit (FTU) and feeder state are taken as decision variables. Based on the theory of minimum fault diagnosis set, a fault location model considering the missing and false information of the FTU is established. In order to solve the problem of complex modeling and a limited algorithm in a large-scale distribution network fault section by logic operation, a switch function based on an algebraic relation operation is constructed, and a logic relation linearization strategy is proposed to linearize the model constraint. In order to ensure the global optimization of the solution, the conventional optimization method with good numerical stability is adopted in the model. The simulation results show that the proposed method is accurate, efficient and fault-tolerant. It can obtain the status information of missing and false alarms of specific nodes while correctly locating the feeder section. This is suitable for a large-scale distribution network with multi node over-current information distortion. This work is supported by Youth Science Fund of National Natural Science Foundation of China (No. 51807114).