Large-scale field studies on the ecological effects of aerial forest spraying often face methodological challenges, such as insufficient funding, difficult logistics, and legal obstacles. The resulting routine use of underpowered designs could lead to a systematic underestimation of insecticide effects on nontarget arthropod communities. We tested the use of an Unmanned Aerial Vehicles (UAVs) for experimental insecticide applications at tree level to increase replication in cost-efficient way. We assessed the effects of two forestry insecticides, diflubenzuron (DFB) and tebufenozide (TBF), on the oak defoliator, Thaumetopoea processionea (Linnaeus) (Lepidoptera: Thaumetopoeidae), and on nontarget, tree-living Lepidoptera. Individual trees were sprayed with either insecticide or left unsprayed, in a fully factorial design involving 60 trees. Caterpillars fallen from tree crowns were sampled as a measure of mortality, while caterpillar feeding activity was monitored by collecting frass droppings. Both DFB and TBF led to greater mortality of T. processionea and lower Lepidoptera feeding activity than control levels. TBF caused measurable mortality in nontarget groups, affecting Macrolepidoptera more strongly than Microlepidoptera, while there was no significant side effect of DFB. The high treatment efficacy against the target pest indicates that UAV technology is well-suited for the application of insecticide in forests. We detected distinct responses to different insecticides among nontarget groups and suggest there is an influence of application timing and biological traits in these differences, emphasizing the need for more ecologically orientated risk assessment. UAV-supported designs can be used to link laboratory bioassays and large-scale experiments, allowing for more comprehensive assessments of insecticide effects in forest ecosystems.