The gas turbine rotor system operates in harsh environments of high temperature, high speed, strong vibration and high stress. Inevitably, various failures will occur, leading to catastrophic consequences. Accurate fault warning and diagnosis are the key to guarantee stable operation of gas turbines. In order to obtain more fault information and comprehensively understand fault characteristics, the cooperative monitoring of multiple types of sensors is an effective means. In response to the redundancy, uncertainty and conflict caused by multi source information fusion, this paper conducted a comparative experiment on the multi source information characteristics of typical faults in gas turbine rod-fastened rotor systems. Firstly, different types and degrees of single type faults and coupled faults were tested on a gas turbine rod-fastened rotor simulation test bed. Fault information was collected using vibration, acoustic emission, and microphone sensors. Then, the sensitivity of each type of sensor to the fault characteristics was comparatively analyzed from the perspectives of time domain and frequency domain, respectively. The experiments show that the targeted analysis of the information collected by different sensors can effectively identify different types and degrees of single fault and coupling fault. This paper preliminarily screened the characteristic information of various types of sensors required for the multi source information fusion fault diagnosis of gas turbine rod-fastened rotor system. It provides support for the realization of accurate fault warning and diagnosis under the condition of cooperative monitoring by multiple types of sensors.