The goal of this work is to study the influence of the structured mesoporous silica (SMS), used as scaffolds for tissue engineering, on the encapsulation and on the release of a model bone morphogenetic proteins (BMP), i.e. Soybean Trypsin Inhibitor (STI). First, SBA-15 silica samples were synthesized using tetraethyl orthosilicate (TEOS) as silica precursor in the presence of P123 as the surfactant in an acidic medium. Three homemade high-pressure stainless steel reactors (with smaller Teflon reactors inside), each equipped with a temperature probe, were used for the syntheses. In addition, a swelling agent was added to increase the pore size. Indeed, the common pore size for those mesostructured materials lies between 6 nm and 8 nm, while the STI has an average radius of 4 nm. The reagent addition sequence was also investigated, by playing on the time of addition of the swelling agent. It was observed that when the swelling agent was added with TEOS, an ordered mesoporous silica was produced. On the contrary, if the swelling agent was added during the surfactant dissolution step, it resulted in an unstructured yet still mesoporous silica. Because of the high affinity of STI for hydrophobic surfaces, SMS were also produced using silica precursors containing phenylene bridging groups (1,4-bis(triethoxysilyl)benzene, BTEB). The BTEB samples exhibited two 2D-hexagonal structures with different wall thicknesses. Synthesized materials were further characterized by infrared spectroscopy (FTIR), thermogravimetric analysis (TGA), nitrogen sorption at 77 K, transmission electron microscopy (TEM), and small-angle X-ray scattering (SAXS). The influence of both the silica composition and structure on STI encapsulation/release was studied. For samples synthesized with TEOS, the unstructured sample showed a higher protein loading and a higher protein release, which could be explained by a difference in the pore interconnectivity within the sample. In fact, a fast release of STI was observed during the first 24 h. Afterwards, the STI release slowed down and seemed to reach a plateau. On the opposite, the structured sample showed a steady release over time. Finally, the sample synthesized with BTEB did not show a significant release over the same period of time. This led us to the conclusion that the hydrophobicity of the silica surface plays a major role on the protein encapsulation and its release rate.