The frequency characteristics of the Daubechies N wavelet series are analyzed. The characteristics of
the frequency band and the influences of the frequency band energy leakage (FBEL) of wavelet are
expounded. The smoothness degree and the distinguishing frequency capability of the Daubechies
wavelet series, and the frequency band energy (FBE) of the signal increase conformably with the
increase of the parameter N. However, the degree of the FBEL decreases with the increase of the
parameter N. An adjacent frequency bands energy increments comparison method (AFBEICM) is
put forward, and this method accords with the energy change rules of the adjacent frequency bands.
With the AFBEICM the interference of the FBEL is eliminated, and the feature extraction is
optimized. In the basis of the AFBEICM, the frequency bands of the wavelet analysis are subdivided
by a fuzzy processing method. It is the basis of the fuzzy decision of the system.ns occur.