Protein engineering endeavours to design new peptides and proteins or to change the structural and/or functional characteristics of existing ones for specific purposes, opening the way for the development of new drugs. This text develops the theoretical formulation for the methods used in computer-assisted modelling and predictions in protein engineering, starting with the basic concepts and proceeding to more sophisticated methods, such as Monte Carlo and molecular dynamics. An evaluation of the approximations inherent to the simulations allows the reader to obtain a perspective of the possible deficiencies and difficulties and approach the task with realistic expectations. Examples from the authors' laboratories are also included.