Thickness is an important influencing factor in liquid non-destructive testing (NDT). Wavelet transform as an important signal processing means has many advantages such as dynamic error threshold determination and multi-scale wavelet transform de-noising. The obtained data are performed multi-level wavelet decomposition and threshold processing of soft threshold through constructing a NDT platform. Comparing with traditional filtering method, the advantages of wavelet transform used in laser NDT are verified. Experimental results show that wavelet transform in laser NDT has better denoising effect and has great significance for liquid NDT on public.