Audio signals are often contaminated by background environment noise from audio equipments. Audio denoising aims to attenuating the noise while retaining the underlying signals. The focal point of this thesis is to use digital signal processing techniques based on wavelet transform to reduce the noise from the signal. In this research three types of wavelet transform (Haar, D4 and D6 wavelet transforms) and different thresholding criteria have been investigated to truncate the noise from two types of signal (high and low amplitude audio signals). The denoising results have been analyzed and compared with the original signal in order to find out the best for audio denoising setups. Furthermore, the performance of all considered methods had been evaluated. The results shown in this thesis indicate that the best denoising results occur when applying scanned thresholding mechanism without making signal framing, and when all coefficients of the detail subbands are thresholded. Among the tested wavelet transforms, the D6 wavelet transform leads to better denoising results. The best denoising results according to the objective measure mean square error (MSE) occurred when using supersoft thresholding, while the best denoising results according to subjective test is when using semisoft thresholding.