The data collection method of turbine blade vibration test was studied, steam turbine blade vibration test was carried out by using the grating fiber and resistance strain gauge measurement method. Rotor running experiment was finished by simulating steam turbine operation state, the experimental program was put forward. Vibration signal data set was collected on the basis of experimental. Apriori algorithm was applied to analyze the data sets by using WEKA platform. Steam turbine rotor vulnerable parts was located in this state, and the blade vibration associated with association rules was found. Research result show the turbine blade vibration problems can be discovered by applying Apriori algorithm, it has the important effect of turbine operational vibration optimization.