Predation is a prime selective force shaping prey behavior. Investment in anti-predator behavior is traded-off against time and energy for other fitness-enhancing activities such as foraging or reproduction. To optimize this benefit/cost trade-off, prey should be able to innately and/or by experience modulate their behavior to the level of predation risk. Here, we assessed learned predation risk management in the herbivorous two-spotted spider mite Tetranychus urticae. We exposed spider mites coming from benign (naïve) or high immediate predation risk (experienced) environments to latent and/or no risk and assessed their site choice, activity and oviposition. Benign environments were characterized by the absence of any predator cues, high immediate risk environments by killed spider mites, physical presence of the predatory mite Phytoseiulus persimilis and associated chemosensory traces left on the surface, and latent risk environments by only predator traces. In the no-choice experiment both naïve and experienced spider mites laid their first egg later on leaves with than without predator traces. Irrespective of predator traces presence/absence, experienced mites laid their first egg earlier than naïve ones did. Naïve spider mites were more active, indicating higher restlessness, and laid fewer eggs on leaves with predator traces, whereas experienced mites were less active and laid similar numbers of eggs on leaves with and without predator traces. In the choice experiment both naïve and experienced spider mites preferentially resided and oviposited on leaves without predator traces but experienced mites were less active than naïve ones. Overall, our study suggests that spider mites experienced with high predation risk behave bolder under latent risk than naïve spider mites. Since predator traces alone do not indicate immediate risk, we argue that the attenuated anti-predator response of experienced spider mites represents adaptive learned risk management.