Several competing hypotheses have been put forward to explain why females of many species mate preferentially with males possessing the most conspicuous signals (e.g., ornaments, displays, or songs). We performed a laboratory experiment using two species of poison frogs, Dendrobates leucomelas and Epipedobates tricolor, to test the hypothesis that male calling performance is an honest indicator of parental quality. Our analyses are based on data from behavioral observations of mating activities of captive-reared individuals (and their offspring) that were housed in terraria for four consecutive breeding seasons. Male mating success increased with male calling rate and chirp duration in both species, suggesting that females preferred males with more elaborate calls. Because calling performance improved with age in D. leucomelas, female poison frogs that prefer males with more elaborate calls in the wild may end up mating with older males that have already proven their ability to survive. Females that mated with good callers obtained higher quality offspring. Eggs fertilized by males with high calling rates and long chirp durations had higher hatching success and produced tadpoles that were more likely to metamorphose into surviving frogs. As a consequence, females that mated with males with high calling performance obtained more surviving offspring per egg, compared to females that mated with poor callers. Collectively, our findings comply with the notion that female poison frogs prefer to mate with good callers because calling performance is a reliable predictor of offspring quality. The possible influence of maternal allocation and reasons for the strong effect size compared to previous studies are discussed.