The vibration fault, one of the common faults in the steam turbine generator unit, brings great damage to the production and the running process. It is well known that the information entropy is to describe the degree of indeterminacy of the system, so the information entropy can be used to measure the vibration condition of the unit. Despite its efficiency, one kind of information entropy is just enabled to identify certain part of the faults. In order to make up for this limitation, based on the faulty signals collected from the rotor test platform, the grey correlation analysis was studied for vibration fault diagnosis of steam turbine shafting in this paper. The reference faulty matrix and the calculation model of grey correlation degree was established based on three kinds of information entropy. The analysis shows that grey correlation analysis is a useful method for fault diagnosis of shafting and can be used as a quantitative index for fault diagnosis. The vibration fault, one of the common faults in the steam turbine generator unit, brings great damage to the production and the running process. It is well known that the information entropy is to describe the degree of indeterminacy of the system, so the information entropy can be used to measure the vibration condition of the unit. Despite the efficiency, one kind of information entropy is just enabled to identify certain part of the faults. test platform, the gray correlation analysis was studied for vibration fault diagnosis of steam turbine shafting in this paper. The reference faulty matrix and the calculation model of gray correlation degree was established based on three kinds of information entropy. The analysis shows that gray correlation analysis is a useful method for fault diagnosis of shafting and can be used as a quantitative index for fault diagnosis.