Random processes and disorder structures that are self-similar on certain length and time scales are very common in nature. The major achievement in recent years that has increased our understanding of structural disorder and its formation by random processes is the fractal geometry. Basically, the fractal theory is based on the self-similarity concept and the central problem is linked to the legitimacy of applying self-similarity. The mathematical concept of fractality, which is valid for an infinite range of similitude, must be restricted to a finite range of similitude in the real physical systems. The presence of these limits complicates the analysis of the experimental data and much of the controversy over the values of the fractal dimension stems from such difficulties. In this article, we review how to characterize the structural heterogeneity of porous materials with the aid of fractal theory. Also discussed are fractal approach to both the heterogeneous catalysis processes and the environmental pollution problems. The emphasis is placed on the applications and limitations of fractal analysis to the above-mentioned systems.