Lethal interactions can shape ecosystem structure, and consequently understanding their causes is ecologically relevant. To improve both empirical and theoretical knowledge on superpredation (i.e. predation on high‐order predators), we studied an eagle owl population, including its main prey and mesopredators, and then we crossed these results with existing theories to provide a reasoning framework. We fitted our field data into four main causes explaining lethal interactions: food stress, opportunistic superpredation, removal of a competitor, and removal of a potential threat. Empirically, superpredation seemed to be mostly determined by the combination of the food‐stress and opportunistic‐superpredation hypotheses, which highlights the complexity of the factors triggering superpredation. Therefore, besides being a response to lower food availability, superpredation may also represent an effective mechanism to remove potential predators and/or competitors, either intentionally or not. Our theoretical framework focused on the decision‐making process in superpredation, considering four inter‐related stages: encountering; attacking; and capturing a mesopredator; as well as consuming a mesopredator once killed. Superpredation almost certainly results from a complex process of decision‐making, accounting for costs and benefits assessed moment‐to‐moment and for each mesopredator individual. It is time to build bridges between theoretical and empirical studies to further understand the mechanisms driving complex interactions among top predators and mesopredators.