In order to improve the denoising effect and the edge detail protection of traditional multi-scale transform filtering in the fringe pattern of electronic speckle pattern interferometry (ESPI), an improved nonsubsampling contourlet transform(NSCT) filtering algorithm was proposed. Discrete stationary wavelet transform and NSCT model were adopted with nonlinear diffusion and improved fast non local mean filtering algorithm. The algorithm was applied to the quantitative analysis of simulated and experimental ESPI fringe pattern filtering results. After theoretical analysis and experimental verification, the results show that, the minimum speckle index of the algorithm in this paper is 0.41121, 0.38043 and 0.35362 respectively in simulated and experimental ESPI fringe pattern. The peak signal-to-noise ratio is the largest. This algorithm improves the ability of denoising and recovers the details of stripes.The results lay the foundation for the application of multi-scale transform filtering in ESPI fringe pattern.