In order to diagnose electrical contact fault of Gas Insulated Switchgear (GIS), the monitoring method for electrical contact state of GIS using the vibration signal is presented. The mechanism of contact vibration is analyzed. The model for coupling multiple physical fields in GIS has been developed. The vibration signals of GIS contact with different state under different operating currents are simulated. The GIS prototype and laboratory test platform are developed. The online monitoring experiment is implemented in the substation. The findings indicate that, at an equivalent current load, when GIS is normal state, the maximum amplitude in the frequency domain is 0.014 m/s2 at 100 Hz, and its high-frequency signal component is 0.003 m/s2 at 800 Hz; when GIS is abnormal state, the maximum amplitude in the frequency domain is 0.056 m/s2 at 800 Hz, and the amplitude of 100 Hz is 0.014 m/s2. Further, the vibration information is extracted by proposed wavelet packet method. Aiming at normal contact state of GIS, the energy entropy in the first and eighth bands are 0.601 and 0.023. Aiming at abnormal contact state of GIS, the wavelet energy entropy of the signal in 1st and 8th bands are 0.041 and 0.142. The conclusions show that, the monitoring method based on vibration information can accurately identify the electrical contact state of GIS.