Abstract: In order to effectively suppress the end effect in empirical mode decomposition (EMD), this paper proposes a processing method based on the combination of mathematical model and signal analysis method which the end effect suppression method of EMD based on grey prediction model. Firstly, the gray prediction model is used to extend the two ends of the original signal to produce limited extreme points. A rectangular window is added to the middle part of the extended signal to keep the original signal unchanged, and a Hanning window is added to the extended part to make the envelope fitting smoother. Then EMD is used to decompose the processed signal sequence to obtain the intrinsic mode components (IMF). Finally, the extended part of the IMF is discarded, so as to reduce the noise It can effectively suppress the endpoint effect. By processing and analyzing the simulation signal and the experimental signal, taking the root mean square error method and the similarity coefficient method as the quantitative evaluation index, the IMF component before and after processing is compared with the component of the original signal, which shows that the improved method in this paper can achieve the purpose of suppressing the endpoint effect.