The purpose of this research is to establish a methodology to describe and detect the cropping system diversity in the Maghreb oases from an interdisciplinary approach. The method calls for concepts originating from semiotics and systemics domains together with multivariate statistics on independent sets of observed data. First, different types of cropping systems have been extracted from multiple correspondence analysis of surveyed agronomic variables collected in situ ( direct observations and measures and/or obtained from interviews). Then, a discriminant factorial analysis was used to assess the quality of the predictive associations between these types and datasets covering a range of characteristics from geometric and radiometric indicators accounting for culture plots physiognomy, the data collection of which did not require here any field survey. Whether to detect the how, the where or the what of the cropping system, the method appears to reach an encouraging discriminating capacity. This pilot study was carried out in the Nefta Oasis, Djerid, Tunisia. The generic aspect of the approach and its potential are discussed, as contribution to population-environment observatories for Oasis society and territorial managers in arid zones.