Microsimulation models play a key role in ex ante evaluations of public policies in that they offer the advantage, by their very nature, of addressing not only the average effects of such policies, but also the dispersion of their consequences. This issue focuses more particularly on the methodological aspects of constructing such models. The fields covered here are those in which microsimulation is most generally used: taxation and redistribution, pensions, dependency and health, etc. Following the usual typology, the models presented here are either static or dynamic. The static models seek to reproduce the impact of policies on individual situations at a given date. The dynamic models, meanwhile, propose to track individual trajectories by artificially generating the main events in the life cycle, such as the end of studies, forming a couple and possible periods of unemployment, etc. This preface provides a reminder of the advantages and drawbacks of these models when compared to more aggregate models. It then goes on to present the methods that can be used to minimise the stochastic character of the results they provide. Finally, it presents the types of databases on which these models can be based and the problems posed by the calibration of their parameters and their validation.