The use of intermolecular perturbation theory to rationalise intermolecular forces and to provide a framework for concepts of intermolecular interactions has been routine for many years. It has long been assumed, however, that the route to accurate calculations of intermolecular potentials lies throughabinitiosupermolecule calculations, and it is only in relatively recent times that this view has been challenged. It is now clear that accurate intermolecular potentials can be constructed for regions where the molecules do not overlap by assembling the components described by long‐range intermolecular perturbation theory−the electrostatic, induction and dispersion terms−though perturbation theory itself does not necessarily provide the best way to calculate them. At short range, the repulsion must be added, and the long‐range terms modified to account for the effects of overlap. Here too, perturbation theory provides a useful route to accurate numbers. At present, the resulting potentials are usually too elaborate for use in simulations on large molecules. However an understanding of accurate potentials is the prerequisite to the construction of simpler models that make large calculations feasible without sacrificing accuracy unnecessarily.