Fractal theory was applied to characterize pore structure. By image processing and fractal analysis for SEM images of sintered stainless steel fiber porous materials, the influence of image resolution, magnification, and the threshold value from grayscale image to binary image on fractal analysis results of pore structure is investigated and their qualitative relationship with fractal dimension was also studied. The results indicate that sintered porous fiber materials have typical fractal nature; the image resolution, magnification and threshold value obviously impact fractal dimension: the larger the difference between original and applied resolution is, or the smaller the magnification is, the larger the fractal dimension becomes; there is a best threshold value for SEM image conversion from grayscale image to binary image, and with the threshold value increasing, the fractal dimension increases and then decreased. The mathematical relations between fractal dimension of pore structure and porosity as well as permeability of porous materials are also discussed. It is feasible to predict the porosity and permeability through the fractal dimension, surface porosity, pore size and other parameters reasonably and the predicted results are in good agreement with the measured results.