A study revealed predator polymorphisms can be maintained by prey learning through negative frequency-dependent selection (NFDS) in which fitness of rare types is greater than common types. Although past studies have revealed that prey polymorphisms can change the population dynamics of predators and prey, there is currently no theory, to our knowledge, about how predator polymorphisms affect predator-prey population dynamics. In the field, comparative observations of snake-lizard predator-prey relationships on two islands indicated that the period of predator-prey cycles on the island with dimorphic snake predators was longer than that of the island with monomorphic predators. To account for this empirical finding, we constructed two mathematical models of predator-prey interactions with one assuming dimorphic predators and the other monomorphic. Comparing the dynamics of these models found that prey learning of alternative types of dimorphic predators increases the period of predator-prey cycles. Moreover, parameter-dependence analysis found that a parameter determining the attack rate of the rare type have strong influences on the period of predator-prey cycles. Our results suggest that NFDS on dimorphic predators through prey learning can increase the period of predator-prey cycles.