Aiming at the problem that the unprocessed multi-source alarm information may cause the fault processing time to increase after the failure of the power system, an artificial intelligence-based fault diagnosis method for the transmission line is proposed. It improves the Petri net model of transmission line fault diagnosis, presets the transition threshold in the established fuzzy Petri net model, and uses the reverse search strategy to reduce the fuzzy Petri net, decrease the inference scale and improve the fault diagnosis efficiency. The proposed method can quickly analyze and classify fault information, greatly reduce the workload of dispatchers processing information, improve the automation level of transmission lines, and ensure the safe and stable operation of power systems. The effectiveness of the proposed method is verified by an example. This work is supported by National Natural Youth Science Foundation of China (No. 51407113).