Pest management strategy aims to minimize the adverse impact of pests on the crop. Basically its building requires a balance of knowledge about farming practices, crops, pests, soil, and their interactions in regard to the landscape structure. New technologies to develop ecologically based pest management require additional knowledge about functional biodiversity in a given agrosystem. In the case of the conservation biological control, knowledge about the local landscape, i.e. natural vegetation and mosaic of crops, and its management also have to be considered to ensure the survival of natural enemies. Thus, many interactions in population dynamics have to be investigated simultaneously, with the possibility of conflicting interactions. In order to identify potential conflicts and provide sound evidence, interactions affecting sugarcane health were implemented in Cogui software for modeling. Using this software, all these interactions were represented as semantic graph, and their automatic combination reveals pest management strategy for a specific location. The interactions concerned food webs regarding different communities such as Lepidoptera, Coleoptera, Hemiptera, Hymenoptera, entomopathogens etc. and the impact of agricultural practices comprising IPM strategies, soil preparation, cultivar, irrigation, fertilization, harvest etc. on the agro-ecosystem. Using this knowledge-based system can help analyses pest problems and propose IPM solutions through thorough comprehension of the pest-natural enemy interactions in a complex landscape structure. (Texte intégral)