With advances in technology, biopharmaceuticals are increasingly used to treat a wide range of diseases, with particular advantages in targetted and tailored treatment compared to conventional drugs. The complexity of biological structures, which lend to their specificity, also cause an inherent sensitivity to their environment. It is therefore a bottleneck in biologic production, storage and delivery to conserve their intricate and distinctive structures. Because therapeutic proteins are engineered primarily to combat disease, stability is often a secondary consideration, which can lead to difficulties in prospective clinical trials. Instabilities can result in more favourable interactions between proteins rather than the solvent around them, promoting formation of oligomeric structures which can aggregate into larger insoluble particles. One strategy to tune protein stability is to modify the solution chemistry by addition of small molecules termed excipients. Each excipient has a function in stabilising the solution to ensure proteins remain soluble and stable. Despite the ubiquitous presence of excipients in protein formulations, the mechanism by which they influence solution stability is not often clear, particularly because multiple and transient interactions with anisotropic proteins are difficult to isolate and characterise. One approach to investigate protein-excipient interactions in greater detail, is to simulate and model molecules using computational techniques. The advantage of computational studies is the scope to examine atomistic details of solutions and therefore to hypothesise mechanisms which support experimental observations. We approach computational studies of protein-excipient interactions by two methods. Firstly, we implement and further develop the energy and entropy theory originated within our research group to assess molecular level stability of constituents in solutions, applied firstly to simpler liquids up to protein-excipient solutions. We present methods to discretise solvent around solutes in a local manner such that the influence of nearby solutes can be characterised. Secondly, we develop an implicit model which predicts how charged excipients influence the overall charge on proteins upon binding, which is an experimental characteristic used to assess possible protein destabilisation in solution. The proteins studied here are small, globular in shape and frequently used in both experimental and computational studies as model systems to characterise solution behaviour. The hope is for our methods to aid in a better understanding of molecular-level solution thermodynamics and the prediction of protein characteristics in solutions.