Since early 1990s, the development of footprint models has been rapid with presently four different approaches being available: (i) analytical models, (ii) Lagrangian stochastic particle dispersion models, (iii) large-eddy simulations, and (iv) closure models. Parameterizations of some of these approaches have been developed, simplifying the original algorithms for use in practical applications. The paper provides a review of the footprint modelling. It also discusses our present understanding of the theoretical background, the most successful modelling approaches, as well as the usage and benefits of the footprint concept as it relates to flux measurements. There has recently been a trend emerging in modelling the behavior of the footprint functions using a less idealized, more realistic description of inhomogeneities, vegetation structure and topography, ultimately for reactive compounds. The estimation of footprints for application in the real world, complete with a multitude of interesting gaseous and particulate substances, remains a complex problem.