Predation is a key ecological interaction affecting populations and communities. Climate warming can modify this interaction both directly by the kinetic effects of temperature on biological rates and indirectly through integrated behavioural and physiological responses of the predators and prey. Temperature dependence of predation rates can further be altered by predator-induced plasticity of prey locomotor activity, but empirical data about this effect are lacking. We propose a general framework to understand the influence of predator-induced developmental plasticity on behavioural thermal reaction norms in prey and their consequences for predator-prey dynamics. Using a mesocosm experiment with dragonfly larvae (predator) and newt larvae (prey), we tested if the predator-induced plasticity alters the elevation or the slope of the thermal reaction norms for locomotor activity metrics in prey. We also estimated the joint predator-prey thermal response in mean locomotor speed, which determines prey encounter rate, and modelled the effect of both phenomena on predator-prey population dynamics. Thermal reaction norms for locomotor activity in prey were affected by predation risk cues but with minor influence on the joint predator-prey behavioural response. We found that predation risk cues significantly decreased the intercept of thermal reaction norm for total activity rate (i.e. all body movements) but not the other locomotor activity metrics in the prey, and that prey locomotor activity rate and locomotor speed increased with prey density. Temperature had opposite effects on the mean relative speed of predator and prey as individual speed increased with temperature in predators but decreased in prey. This led to a negligible effect of body temperature on predicted prey encounter rates and predator-prey dynamics. The behavioural component of predator-prey interaction varied much more between individuals than with temperature and the presence of predation risk cues in our system. We conclude that within-population variation in locomotor activity can buffer the influence of body temperature and predation risk cues on predator-prey interactions, and further research should focus on the magnitude and sources of behavioural variation in interacting species to predict the impact of climate change on predator-prey interactions and food web dynamics.