Abstract: For the unstable Gaussian noise of the CCD image, the conventional method of Fourier transform can not eliminate it,but the wavelet transform is the powerful tools of noise elimination,since the wavelet transform in both the time domain and frequency domain has the excellent localization features, and the arbitrary details of analyzed objects can be focused with gradually detailed time domain or frequency domain step at high frequency. Based on the analysis for the CCD image noise,the noise reduction and edge detection is realized with the method of wavelet transform and histogram division. The results show that the clear image edge can be made out after selecting a proper threshold with the histogram division method.