Parental investment theory proposes two non-mutually exclusive hypotheses to explain variation in anti-predator behaviour in relation to the age of offspring: the “reproductive value of offspring” hypothesis and the “harm to offspring” hypothesis. The relative importance of the two factors underlying the hypotheses, reproductive value and harm, may change depending on environmental conditions such as food availability. To test the relative importance of the two hypotheses under different food conditions, we conducted a supplementary feeding experiment in red kite (Milvus milvus) breeding pairs and used a live eagle owl (Bubo bubo) as decoy nest predator to trigger anti-predator behaviour. We used time-to-capture in mist nets mounted next to the decoy predator as proxy for mobbing intensity. Under natural food conditions we found a nearly constant mobbing intensity throughout the entire nestling period. However, under food-enhanced conditions mobbing intensity was reduced in parents with young nestlings and increased in parents with old nestlings. These results suggest greater importance of the “reproductive value of offspring” hypothesis in situations of favourable food availability. Moreover, mobbing intensity depended on brood size and weather conditions. The results suggest that parental anti-predator investment increases with the reproductive value of the brood under favourable breeding conditions, but that this pattern is adjusted to the current context, including the vulnerability of the brood and environmental conditions.