In order to ensure that small diameter steel pipes with thick wall have high intensity and high quality, ultrasonic immersion method with focusing probe was used to detect the flaw of the small-diameter steel pipes with thick wall. In practice, the echoes are often corrupted with external noise or internal noise, therefore, it is necessary to reduce the noise and to enhance the SNR of ultrasonic signals. A technique for improving the SNR of ultrasonic signals using wavelet transform is presented. In this method, WT, consider as one band-pass filter, is used to remove the noises. The performance of this technique has been verified by experimental, which is done by using a series of flaw ultrasonic echoes obtained from a specimen of the small-diameter steel pipes with thick wall. In particular we have found the processing of the ultrasonic signals using wavelet transform extremely useful for noise reduction. After processing, the SNR of ultrasonic signals are enhanced substantially. All experimental results show that this technique is effective for removing the white noise from the ultrasonic signals.