Abstract: Using the signal processing system of virtual ultrasonic testing which was designed by ourselves, the ultrasonic testing signal of pipe with rough surface has been processed based on the Fourier transform and Wavelet transform respectively to solve the problem of low signal-to-noise ratio. The signal processing work that based on the Fourier transform is mainly refers to the spectrum analysis and FIR filter. The signal processing work based on the Wavelet transform is mainly refers to the time-frequency analysis and Wavelet denoising. By comparing the experimental results, it has shown that the Wavelet transform method was better than the Fourier transform when processing the ultrasonic testing signal of pipe with rough surface.