Abstract: A new denoising algorithm based on adaptive threshold is proposed.First,the threshold function is continuous and has infinite rank continuous derivative,fully reflects the energy distribution of wavelet coefficients. What's more,the function expression is simple and easy to calculate,so it is suitable for various mathematical processing.Secondly,the correlation and procedural of wavelet coefficients in decomposition process is considered to the threshold's selection,which will reduce the false positive probability of noises and will be more practical. Simulation results show that the new algorithm not only has smaller computational cost than traditional algorithms, and achieves a higher PSNR and smaller MSE.It can remove the noise more effectively.