A method for ECG de-noising based on Hilbert-Huang Transform has been proposed.ECG was analyzed by the Hilbert spectrum of the IMF

which was produced by the Empirical mode decomposition method.The three main noises were respectively removed according to the characteristics of noises.Power line interference and high frequency noises were mainly mixed into the low IMF of ECG

while baseline wander noises were mainly mixed into the high IMF.The morphological filter method based on adaptive threshold was used in the low IMF for de-noising

while baseline wander was estimated by the smooth filter in the high IMF of ECG.The results of simulation experiment and practical application demonstrate that the method proposed in this paper performs better than wavelet de-noising method significantly

not only in suppressing the three mainly noises effectively

but also in preserving the primary characteristics of ECG signal.