A method and a system for active power distribution network fault diagnosis and location are provided. The method comprises the following steps: S1, dividing a line into N sections; S2, acquiring fundamental component signals of the voltage and current at the start and end of a kth section; S3, establishing a Pi-type four-port network model of the kth section under a non-fault condition to work out a theoretical value; S4, comparing the theoretical value with a measured value; S5, if the comparison results are within the allowable range of error, letting k=k+1, and jumping to S7; S6, if there is at least one comparison result not within the allowable range of error, establishing a relation equation of sequence components of a Pi-type four-port network of the kth section under a fault condition to judge the fault type and fault position, letting k=k+1, and jumping to S7; and S7, jumping to S2 is k<=N, and completing line detection if k=N+1. The system comprises a signal acquisition module and a central processor which are connected through a communication network. The type and position of a fault can be located accurately.