Bit vibration wave is applied to cluster well anti-collision, in order to accurately identify the vibration signals of the drill bit, the HHT method is applied to the bit signal feature extraction, using the EMD decomposition can make bit vibration signal decomposed into intrinsic mode components(IMF), the Kurtosis criterion and Correlation coefficient is used to filter the main impact components and then restructures filtering signal. Marginal spectrum, instantaneous frequency and the three dimensional spectrum , obtained by Hilbert transform, is used to analize and compare to judge wheather the signals of differernt channels come from the same drill bit or not, extract the main public frequency range as the bit vibration signals, the bandpass filter is applied to the channel to extract bit vibration signal,according to the energy of the filtered signal to establish a bit approach to the risk well model, the proposed processing method was confirmed effective though real experiment data ayalysis on the offshore cluster well.