Traditionally, the search for active and selective catalysts involved a rather tedious "hit-and-miss" approach in which hundreds, if not thousands catalysts were tested until the optimal substance was identified. With the advancing theoretical understanding of catalysis and the development of computational power, a new era of rational catalyst design is dawning. This approach, grounded on first principles, is based on new advances in synthesis, characterization and modeling with the ultimate aim of predicting the expected behavior of a catalyst based on chemical composition, molecular structure and morphology. The rational design of a catalyst is a complex process which spans across several levels of scale. In this thesis, the pursuit of a rational design for Pd-based catalysts effective in alkyne hydrogenations is presented. A multi-level integrated approach was thus applied ranging from the nano-scale design of the active sites for a specific reaction, taking into account its structure sensitivity, to the micro-scale design of the supported Pd nanoparticles including metal-additive and metal-support interactions as well as mass and heat transfer phenomena. In order to rationally design a catalyst at a nano-scale, i.e. a catalyst's active site, several methodologies have been hitherto applied, such as the study on single crystals or model catalysts. Here we present the use of metal nanoparticles with tuned sizes (5-30 nm) and shapes (cubes, octahedra and cube-octahedra) prepared via colloidal techniques. These nanoparticles can be tested per se and represent a new generation of model catalysts, complementing single crystal studies, which inherently lack the complexity of industrial catalysis. However, metal nanoparticles, especially those prepared in a controlled manner in order to tailor their shape and size, require the use of stabilizing and/or capping agents capable of directing their growth. These substances can mask the true catalytic behavior of the nanoparticles. Thus, it is of great importance to study the interactions of the active phase with the substances that are in close contact with it, in the so-called meso-scaled level of rational catalyst design. Therefore, the effect of the nature of the stabilizing agent on the catalytic response was studied, as well as the promoting effect of some of these substances. Finally, a methodology was developed capable of eliminating organic stabilizing agents from the surface of nanoparticles without compromising their morphological stability. The knowledge gathered in the first two levels can be applied further to reach the microscale rational catalyst design. In this thesis, a final catalyst consisting on well-defined stabilizer-free supported Pd nanoparticles was used in the hydrogenation of acetylene. This deep study of the catalytic behavior of well-defined Pd catalysts throughout several levels of scale and complexity have given us the tools needed to perform a rational catalyst design for alkyne hydrogenations. Depending on the specific reaction, the active phase can be optimized in terms of the desired activity and selectivity and can be tuned even further with the use of specific additives. Finally, the appropriate support also exerts a promoting effect on the nanoparticles and ensures their anchoring in addition to avoiding heat and mass transfer artifacts.