In order to be able to estimate the deformation in the nature effectively and reduce the estimation error,a method of MEMS inertial deformation estimation is proposed in this paper. Firstly, the signals from MIMU are collected and the data signals are processed by using Kalman filtering to eliminate the disturbance errors.Then aiming at the deficiency of grey model in deformation estimation, the modified gray model combined with Kalman filtering is applied to the deformation estimation to carry out the grey prediction.Example application shows that the combined model is more suitable for the deformation estimation than the pure grey model;it can be used as a new kind of prediction method to apply to deformation monitoring.