Birds and predatory arthropods are often implicated in pest control, but their relative impact and how this is mediated by variation in tree cover requires elucidation. We excluded birds and ground predatory arthropods from rapeseed plants in 2.5. ×. 1. m plots in 26 homegardens in Ethiopia, leaving the same sized control plots. From six groups of plants in bird exclosure and control plots, respectively, three groups were excluded from ground predatory arthropods. Data on leaf damage were surveyed four times at weekly intervals. The tree cover and land-use composition within 100. ×. 100. m surrounding each plots were recorded in the field and from a satellite image within 200 and 500. m buffer zones. The results show that the mean leaf damage was higher on rapeseed plants from which predatory arthropods were excluded than on control plants. However, excluding birds had no or only a weak impact on leaf damage. The mean leaf damage within predatory arthropod exclosures decreased with increasing tree, forest and perennial cover but increased with increasing grazing land cover and annual crop cover, while on control plants it was low across the tree cover variation. This pattern may indicate the presence of a higher density of herbivores on rapeseed plants and also more predatory arthropods (i.e., to control them) in tree-poor homegardens compared to tree-rich homegardens. Hence, tree-poor homegardens in this landscape have sufficient habitat heterogeneity to support natural enemies to deliver significant pest control on rapeseed. Our results show that there was variation in the dynamics of pests and predatory arthropods across the tree cover variation, suggesting changes in landscape composition could affect the pest control services and the outcomes for local farmers.