Permeability is one of the key parameters for quantitatively evaluating groundwater resources and accurately predicting the rates of water inflows into coal mines. This paper presents an efficient method to estimate the macroscopic permeability by using the scanning electron microscopy (SEM) images. A correlation between the microscopic features of sandstone porosity and the macropermeability is approached by an image identification technique. Firstly, the gray images were transformed into the binary images by using the histogram of the entropy method. Then, the Green and Euler distance methods were applied to calculate the length and area of the pores, and the fractal parameters were estimated according to the slit island method. Based on the theory of microscopic seepage flow, the seepage coefficient and permeability were calculated by fractal parameters. Typical water-bearing sandstone samples in the Kailuan coal field area in North China were selected to demonstrate the methodology. SEM microscopic images of nine groups of sandstone samples collected at different depths (from the outcrop to the deep mines) were analyzed. Based on the theoretical model of micropore structures and the fractal theory, the permeabilities were estimated. The results provide insights for understanding the hydraulic properties of the sandstone.