The world wide decline of insects is often associated with loss of natural and semi-natural habitat caused by intensified land-use. Many insects provide important ecosystem services to agriculture, such as pest control or pollination. To efficiently promote insects on remaining semi-natural habitat we need precise knowledge of their requirements to non-crop habitat. This thesis focuses on identifying the most important semi-natural habitats (forest edges, grasslands, and semi-open habitats) for pollinators and natural enemies of crop pests with respect to their food resource requirements. Special attention is given to floral resources and their spatio-temporal distribution in agricultural landscapes. Floral resource maps might get closer at characterizing landscapes the way they are experienced by insects compared to classical habitat maps. Performance of the two map types was compared on the prediction of wild bees and natural enemies that consume nectar and pollen, identifying habitats of special importance in the process. In wild bees, influences of spatio-temporal floral resource availability were analysed as well as habitat preferences of specific groups of bees. Understanding dietary needs of natural enemies of crop pests requires additional knowledge on prey use. To this end, ladybird gut contents have been analysed by means of high-throughput sequencing for insight into aphid prey-use. Results showed, that wild bees were predicted better by floral resource maps compared to classical habitat maps. Forest edge area, as well as floral resources in forest edges had positive effects on abundance and diversity of rare bees and important crop pollinators. Similar patterns were retained for grassland diversity. Especially early floral resources seemed to have positive effects on wild bees. Crops and fruit trees produced a resource pulse in April that exceeded floral resource availability in May and June by tenfold. Most floral resources in forest edges appeared early in the season, with the highest floral density per area. Grasslands provided the lowest amount of floral resources but highest diversity, which was evenly distributed over the season. Despite natural enemies need for floral resources, classical habitat maps performed better at predicting natural enemies of crop pests compared to floral resource maps. Classical habitat maps revealed a positive effect of forest edge habitat on the abundance of pest enemies, which translated into improved aphid control. Results from gut content analysis reveal high portions of pest aphid species and nettle aphids as well as a broader insight into prey spectra retained from ladybirds collected from sticky traps compared to individuals collected by hand. The aphid specific primer designed for this purpose will be helpful for identifying aphid consumption by ladybirds in future studies. Findings of this thesis show the potential of floral resource maps for understanding interactions of wild bees and the landscape but also indicate that natural enemies are limited by other resources. I would like to highlight the positive effects of forest edges for different groups of bees as well as natural enemies and their performance on pest control.