This paper proposes an original numerical strategy in order to assess high frequency coupling in large systems, typically for HIRF applications, in a frequency range in which classical full waves numerical schemes such as FDTD or MoM become tedious and sometimes inappropriate to use. This strategy is based on a logical topological decomposition of the electromagnetic (EM) problem as sub elementary problems which are solved one at a time. An example of application of the whole numerical strategy carried out in the context of the HIRF-SE European project is shown. The test-case is a scale model of an aircraft body; the successive modelling steps applied onto this example are detailed and illustrated in this paper.