The Hilbert Huang Transform (HHT) is an advanced algorithm, but its end effect problem limits its applications severely. A new method to deal with the end effect is proposed in this paper. A new computation parameter is applied to estimating the similarity between two waveforms. Both the shape factor and amplitude factor are considered in the new parameter, and the more suitable end extending waveform can be acquired. The window function is designed and added to the extended data wave, and that can reduce the influence of end effect further. In the simulation experiments, time-frequency spectrums and marginal spectrums are given, and by contrasting the spectrums we can draw the conclusion that this new method to deal with the end effect is feasible and it can achieve better results than the traditional algorithms.