Mine safety has always been the top priority in mine mining operation, during the mining process, the surrounding soil will be accompanied by vibration, when the frequency amplitude of vibration reaches a certain critical point, landslides, debris flows and other disasters often occur. Therefore, it is imperative to monitor the vibration of the soil during the mining process. Traditional measurement and monitoring technology due to large amount of work, low accuracy and other shortcomings, that is, the use of global navigation and positioning system (GNSS) to mine soil vibration continuous monitoring, but GNSS system due to multi-path error and other influences lead to the monitoring data contains noise, in order to eliminate noise, obtain accurate vibration signal for subsequent safety analysis, proposed to use Elmd algorithm to process the original data, by adding Gaussian white noise to the original data in the LMD decomposition, according to the correlation coefficient of noise elimination Compared with the EMD algorithm, the root mean square error and signal-to-noise ratio were used to distinguish the effects of the data processed by the two algorithms, and the effect of the ELMD algorithm in the denoising of GNSS vibration monitoring data was verified.