Crystallization pressure is a phenomenon that occurs when crystals form in confined environments, generating mechanical stress on the surrounding walls. This process can cause significant damage in a variety of materials, including cementitious materials and geomaterials. However, despite its importance, this phenomenon is one of the least understood in porous media mechanics. Its direct measurement is particularly complex and requires pore-scale experiments. Moreover, the underlying physical mechanisms, such as the role of the nanometric liquid films at interfaces, are still poorly understood. Consequently, traditional models, such as Correns’ law for salt crystallization, are not sufficient to account for the diversity of degradation caused by salts observed in real life situations. To better understand crystallization pressure, this project, focusing on sodium chloride crystallization, relies on two main approaches: molecular simulations capable of modeling the nanometric film at interfaces, and microfluidic experiments that reproduce and characterize crystallization at the pore scale. This study aims to identify the mechanisms that control the phenomenon, which could pave the way for strategies to prevent salt-induced degradation. At thermodynamic equilibrium, crystallization pressure arises from the variation in solubility of a crystal subjected to compression. Direct molecular dynamics simulations to determine the solubility of salts are unsuitable, as the time scales associated with dissolution and precipitation exceed the microsecond, at the limit of current computational capabilities. To overcome this limitation, we use alternative approaches. A first method is thermodynamic integration, which enables us to directly estimate the osmotic equilibrium conditions corresponding to crystallization, without any link to time evolution. This method also enables us to quantify the effect of stress on salt solubility, in particular the impact of stress anisotropy, an aspect largely neglected until now. The results obtained enable us to revisit the existing theory of crystallization pressure and extend it to incorporate the influence of stress anisotropy. We also use biased grand canonical Monte Carlo methods to determine the critical pressure threshold beyond which the wetting film, which separates the crystal from the pore surface and enables crystal growth, disappears. This approach provides key insights into the film’s stability under different pressure and temperature conditions. Finally, we have explored a third approach based on a recently proposed time-parallelization algorithm (parareal), which enables us to envisage the direct simulation of crystal growth under pressure in the presence of a supersaturated solution. On the experimental front, we have developed a protocol for controlling the precipitation and growth of salt crystals in microfluidic models. This protocol relies on the use of a fluorescent dye coupled with image analysis to detect deformations induced by crystallization pressure.