It is difficult to measure predation risk for secretive animals such as snakes. The frequency of tail breakage has been suggested as a parameter for measuring predation risk in these reptiles. Consequently, we hypothesise that tail breakage frequency is expected to vary intraspecifically between populations submitted to different numbers of predators, probably, to different predation risk. The colubrid snake Natrix maura has adapted to living on fish farms, which are protected from aerial predators by wire mesh, so that these farms provide researchers with a predator-exclusion scenario. We measured tail breakage frequency in this snake on a fish farm and compared the results with data from three populations in natural settings in the Iberian Peninsula under assumed higher predation risk from a complete community of predators. A logistic regression analysis demonstrated that snake body size and predator exclusion affected the probability of tail breakage, while sex had no effect. We found an ontogenetic increase in the frequency of tail breakage, a general pattern in snakes and, notably, marked differences between populations according to predation risk: snakes from the fish farm showed a low frequency of tail breakage (5.1%), whereas snakes from natural habitats, with five-fold more potential predators on average, demonstrated a higher frequency of tail breakage (18.7%). We conclude that tail breakage frequency is a reliable indicator of predation risk in this aquatic snake.