Predators can affect parasite–host interactions when directly preying on hosts or their parasites. However, predators may also have non-consumptive indirect effects on parasite–host interactions when hosts adjust their behaviour or physiology in response to predator presence. In this study, we examined how chemical cues from a predatory marine crab affect the transmission of a parasitic trematode from its first (periwinkle) to its second (mussel) intermediate host. Laboratory experiments revealed that chemical cues from crabs lead to a threefold increase in the release of trematode cercariae from periwinkles as a result of increased periwinkle activity. This positive effect on transmission was contrasted by a 10-fold reduction in cercarial infection rates in the second intermediate host when we experimentally exposed mussels to cercariae and predator cues. The low infection rates were caused by a substantial reduction in mussel filtration activity in the presence of predator cues, preventing cercariae from entering the mussels. To assess the combined net effect of both processes, we conducted a transmission experiment between infected periwinkles and uninfected mussels. Infection levels of mussels in the treatments with crab cues were sevenfold lower than in mussels without crab chemical cues. This suggests that predation risk effects on mussel susceptibility can counteract the elevated parasite release from first intermediate hosts, with negative net effects on parasite transmission. These experiments highlight that predation risk effects on parasite transmission can have opposing directions at different stages of the parasite's life cycle. Such complex non-consumptive predation risk effects on parasite transmission may constitute an important indirect mechanism affecting prevalence and distribution patterns of parasites in different hosts across their life cycle.