Abstract: In the pipeline industry, the health status of turbine-driven compressor set is very critical to keep the stable, scientific and efficient operation of the company, to ensure the safe transportation, and to reduce the operating costs. Herein, with reference to the advantages and successful experiences of similar platforms at home and abroad, the following research was conducted with scientific and effective means to establish a prognostics health management (PHM) system of gas path in turbine-driven compressor set. Definitely, the performance analysis module of gas path was built through modular modeling, the fault diagnosis of the gas path was performed with the gas path fault diagnosis algorithm based on the accurate non-linear compressor model, and a hybrid algorithm of simulated annealing-particle swarm optimization (SA-PSO) was proposed for fault diagnosis, with the rapidity and accuracy considered simultaneously. In addition, a fault feature degraded parameter identification library was established, and a fault probability analysis model was built based on the fault feature recognition method to obtain the quantified probability indicators for the state measurement of gas path components. The application results show that the established gas path analysis program of turbine-driven compressor set could realize the real-time control of the operation status of gas turbine and timely find and diagnose the faults in operation, which has high application value.