In order to reduce the random errors in the MEMS gyroscope,an adaptive compensation method is proposed based on Empirical Mode Decomposition (EMD) and Variational Mode Decomposition (VMD).Firstly,the signal is decomposed into Multiple Intrinsic Mode Functions (IMFs) by EMD,and all the IMFs are divided into noise-dominated IMFs,noise-information mixed IMFs and information-dominated IMFs by using the similarity of probability density functions and the continuous mean square error.Then,the noise-information mixed IMFs are preliminarily de-noised by using the soft threshold interval method (SIT).The number of effective modes is taken as the number of Band-Limited Intrinsic Mode Functions (BLIMF) decomposed by VMD,and the preliminary denoised signal is further denoised by VMD.The correlation coefficient (CC) between each BLIMF and the original signal is used as the basis for whether discarding the mode as the noise mode or not.In order to validate the effectiveness of the proposed method,gyroscope signal under sectional uniform motion is collected and analyzed.The Root Mean Square Error (RMSE) of the denoised signal is 0.0422 of that of the original signal,which is 68.75% of that of EMD-SIT.