The ultra-high frequency (UHF) partial discharge monitoring device which is now used in power grid has the phenomenon of lacking the accuracy of measurement as well as the outcome reliability. In order to improve the metrical precision of UHF partial discharge monitoring system and analyze the features of its error, the analytical method of dynamic measurement system based on Ensemble Empirical Mode Decomposition is provided in this paper. When mutation or interval exists in the signal component of the original signal, using Empirical Mode Decomposition may lead to modal aliasing phenomenon. Otherwise, the Ensemble Empirical Mode Decomposition overcomes the defect in truncated signal and restrained the modal aliasing phenomenon at a certain extent. By using simulation, the Ensemble Empirical Mode Decomposition can make the error signal adaptively decompose into several simple signal so it brings us some convenient in research and finally validates the availability of our method.