Two evolutionary algorithms are introduced as universal approaches for the identification of water transport characteristics of inorganic porous materials in both absorption and desorption phases. At first, genetic algorithm and genetic programming are applied for the inverse analysis of water content profiles measured in an absorption experiment. A comparison of results with the output of the commonly used Boltzmann-Matano approach shows that the calculated diffusivities can reproduce experimental data with a similarly good or even slightly better accuracy. In the second part of investigations, a water desorption experiment is realized for autoclaved aerated concrete, a typical representative of inorganic porous materials used in the construction sector. The genetic algorithm and genetic programming exhibit an excellent performance also in this case. Both approaches can thus be considered as viable, more universal alternatives to the traditional methods.