In this paper the nonlinear stochastic differential equation is used to generalize an intermittency signal

and the correlation of its 1/

f

noise characteristics under the wavelet bases with different vanishing moments are analyzed.The simulation results show that in the middle band of the frequency spectrum

its power spectrum density (PSD) shows typical 1/

f

noise characteristics

and the relationship of the signal's wavelet transformation coefficient variance and the corresponding wavelet scale is linear under logarithm coordinate.In this frequency band

under some scales the signal's wavelet transformation coefficient correlations decrease with the vanishing moments growing

whi

le in other scales

the correlations increase with the vanishing moment increasement.These results indicate that wavelet transformation does not always bring decorrelation to the intermittency signal.The relationship of the wavelet transformation coefficients' variances and the scale

and the tendency of coefficients' correlations with wavelet vanishing moments changing are analyzed.