By elucidating structure-property relationships of soft matter, molecular simulations can assist in the engineering of materials with desirable properties; e.g., properties that originate in the creation of special types of structural order or the control of phase transitions. In mesophases like liquid crystals and plastic solids, e.g., the material has a structural order that lies in between that of solids and liquids, which often offers a combination of characteristics that makes them attractive to create materials with interesting properties. Motivated by the ability of current synthetic routes to generate tailor-made nanoparticles of tunable sizes, shapes, chemistry, and surface patterning, we focus on colloidal particles that can form ordered phases as the basis of a new physical-chemistry with “super atoms” as building blocks. While molecular engineering often focuses on tuning chemical interactions to control the assembly of building blocks, we focus first on the engineering of entropic interactions as they are often equally crucial but less appreciated. I will illustrate these efforts for the case of suspensions of particles with polyhedral shapes. After reviewing some of the general statistical mechanical principles involved in modeling thermophysical properties of molecular models, I will describe selected methods that can accurately evaluate entropy (and free energy) changes associated with different phases and with the transition between phases. After pointing out some of the outstanding methodological challenges, I will present some of approaches that my group has been developing to address them. Sample results are presented on the prediction of novel entropy-aided self-assembling liquid- and plastic-crystalline phases which are resilient to size polydispersity, and how such phases nucleate. These examples illustrate the potential benefits of a more pro-active approach to harnessing the often overlooked power of entropy.