In order to accurately measure the surface deformation information, a surface deformation measurement method based on Empirical Wavelet Transform (EWT) and kernel probability density was proposed. Firstly, the CCD camera was used to collect the digital speckle pattern interferometry (DSPI) maps, the DSPI maps were decomposed to obtain a series of intrinsic components by EWT. According to the kernel probability density of the decomposed component, an adaptive de-noising method was proposed to extract the components with the deformation information, the components were reconstructed DSPI maps; Finally, using Hilbert method to calculate reconstructed DSPI phase, the DSPI phase before and after deformation was subtracted, and the surface deformation information was obtained by decoupling according to the phase difference. The experimental results showed that the method can effectively measure the object surface deformation, and the accuracy is improved by 4 times compared with the EMD method.