Herbivorous insects cause important yield losses to crops. These insects are mainly managed through insecticides but negative effects of used compounds on the human health and the environment as well as the development of resistant pest populations impose finding alternative strategies to manage them. Breeding plants for enhanced resistance to insects is an attractive strategy. This kind of strategy was already implemented to manage insects but examples of its utilisation remain rare. The main limitation of this strategy is that a large number of genotypes needs to be screened through generally complex phenotyping methods to identify resistances. To circumvent these issues, an approach based on the understanding of plant defence mechanisms and the utilisation of biochemical biomarkers linked to resistance could be interesting. During this PhD, this kind of approach was developed on the oilseed rape (Brassica napus) and one of its main pests, the pollen beetle (Brassicogethes aeneus). Previous studies identified chemical compounds present in the perianth of flower buds and correlated to oilseed rape resistance to pollen beetle. However these studies were carried out in the laboratory and need to be validated in field conditions. Moreover, information on pollen beetle feeding ecology are still lacking while they could help identifying new targets for breeding. The first objective of this PhD is to develop a method allowing to screen resistance to the pollen beetle in the field. This method will allow to confirm the potential biomarkers previously identified and to look for additional biomarkers. The second objective of the present work was to better understand key steps in the interaction between the pest and its host plant to identify potential new target traits for resistance. For this purpose, the importance of food sources present in flowers such as pollen and nectar on pollen beetle development and the factors impacting the feeding pattern of adults on the inflorescence were investigated. We conducted field experiments in two different sites and for two consecutive years and propose a method allowing to screen oilseed rape resistance to pollen beetle in the field. Using this method, we were able to identify genotypes with moderate levels of resistance among a set of 19 genotypes. Previously identified biochemical biomarkers were not correlated with plant resistance in the field but new markers were identified (i.e. quinic acid and arginine). Our field experiments also showed that plant composition is highly variable according to the environment and this variability could affect usefulness of these markers during plant breeding programs. Experiments under controlled conditions also showed that pollen beetle used nectar for feeding but that it did not seem to affect its development. Pollen, on the other hand, seemed to have a more important impact but was not indispensable to pollen beetle development. The study of the pollen beetle feeding behaviour showed that this insect has a surprising feeding pattern on oilseed rape inflorescences and that small buds are more used for feeding than large buds that contain more pollen. It seems that accessibility and to a lesser extend availability of feeding resource explain this pattern and that the perianth has a major role on this preference. These experiments allowed to better understand the pollen beetle feeding ecology and to identify plant traits important for its development. Our work showed that moderate levels of resistance are present in oilseed rape and could be used in breeding programs. Limitations of approaches based on biochemical biomarkers are discussed.