Recent experiments an agricultural ecosystems have shown that resilience is the key factor of their sustainability. Here, we suggest that to assess the resilience of agricultural systems with models, it is necessary to consider the complete food web of the simulated systems and not only the productive crops, soil chemistry and eventually crop parasites as usual plant focused cropping systems models do. We attend to reconcile ecological and agronomical approaches through the concept of food web based cropping system model. Food web models are typically used to understand population dynamics and semmunities structures in many ecosystems e.g. aquatic or soil ecosystems. We specifically present a model aimed at simulating a cultivated field as a foil to the traditional cropping system models. We argue that this concept has different implication for understanding field resilience including the interaction between parasite and non parasite communities, physical and chemical soil characteristics. To this end, we offer contrasting prediction of soil characteristics, crop productivity and parasite populations using the SIMBA model with and without taking into account soil food web structure and interactions. Primary production and most damaging parasites of banana (e.g. plant-parasitic nematodes) are simulated by the standard modules of the SIMRA model. Other trophic levels (i.e. decomposers, primary consumers omnivores. predators) are accounted within a simple food web module based on bottunwup and top-down control of populations. Incorporating the variation in food web structure into cropping systems models parameterized for banana systems allows more precise and more realistic simulations of production and parasite dynamics. These results provide evidence that incorporating food web structure can he useful in the search for more accurate prediction of cropping system resilience and for designing more sustainable cropping systems.. The conceptual framework that we present