With an increasing understanding of the cell at the molecular level, primarily guided by advances in high-throughput “omics” and systems biology, metabolic engineering has become more rational and less reliant on trial and error. A key aspect of present-day metabolic engineering is the ability to reliably construct predictive models of cellular metabolism in silico, often at the systems level, and to use these models to predict possible targets for strain improvement. A number of methods have been developed, based on chemical kinetics and constraint-based modeling techniques such as flux balance analysis, as well as network-based methods. In this chapter, we present an overview of the various in silico methods typically employed in metabolic engineering, with particular emphasis on the various success stories.