The phenomenon of Batesian mimicry, where a palatable animal gains protection against predation by resembling an unpalatable model, has been a core interest of evolutionary biologists for 150 years. An extensive range of studies has focused on revealing mechanistic aspects of mimicry and the evolutionary dynamics of mimicry systems and revealed that protective mimicry is widespread and is important for individual fitness. However, case studies where mimicry theories have been applied to conservation of mimetic species are rare. The theoretical implications of mimicry on conservation include, for example, frequency dependency of predator avoidance learning and preventing human induced mortality. We examined the case of the protected, endangered, non-venomous smooth snake (Coronella austriaca) that mimics the non-protected venomous viper (Vipera berus), both of which occur in the Åland archipelago, Finland. To quantify the added predation risk on smooth snakes caused by the rarity of vipers, we calculated risk estimates from experimental data. Resemblance of vipers enhances survival of smooth snakes against predation because many predators avoid touching venomous vipers. Mimetic resemblance is however disadvantageous against human predators, who kill vipers and accidentally kill endangered smooth snakes. Moreover, we found that the effective population size of the adders in Åland is very low relative to its smooth snake mimic. Because Batesian mimicry is advantageous for the mimic only if model species exist in sufficiently high numbers, it is likely that the conservation program for smooth snakes will fail if adders continue to be destroyed. Understanding the population consequences of mimetic species may be crucial to the success of endangered species conservation. We suggest that when a mimic requires protection, conservation planning should not ignore the model species even if it is not itself endangered.