The aim of the present study is to examine whether individuals have the memory advantage for threatening information in the natural environment. To further explore the psychological mechanism of this memory advantage effect by adding threatening and non-threatening scenarios and analyzing whether the memory advantage for threatening information still exists. Threatening information (snake) and non-threatening information (turtle) in the natural environment were selected to be the research materials, which included words and pictures respectively. Specifically, 36 subordinate classification vocabulary of snakes and turtles with middle-level memory difficulties were selected respectively. Because of the fear emotion aroused by real snake pictures may directly exert effect on the memory, 36 cartoon images of snakes and turtles were selected respectively with the medium degree of fear and novelty. Different Scenarios were added in experiment 2 with different instructions. Threatening scenario instruction was like this: “please imagine you are trapped in a prairie in ancient times, and there are all kinds of animal inhabited, with beasts howling and fighting voices in the distance. Some fierce animals are likely to appear around you at any time.” The instruction of non-threatening scenario was: “please imagine you are walking in the zoo now, you can see animals in cages friendly greeting to you , beasts are not as aggressive as those in wild areas, and some well-trained fierce animals are showing skill performance comically.” The results demonstrate that: (1) the recognition memory of snakes which representing threatening information is significantly superior to that of the turtles which representing non-threatening information, no matter the experimental material presenting in vocabulary or image. The result suggests that there is a significant memory advantage for both the threatening vocabulary and image information in the natural environment.; (2) there is a significant difference in recognition accuracy between threatening vocabulary and threatening images information, suggesting that memory of image information is performed better than that of the vocabulary information; (3) when in threatening scenario and non-threatening scenario in experiment 2, the recognition accuracy of threatening information is significantly higher than that of the non-threatening information as what’s found experiment 1. But the recognition accuracy is no significantly different between the threatening scenario and the non-threatening scenario in the natural environment; (4) specifically, when in threatening scenario, the recognition accuracy of threatening words and pictures are both significantly higher than that of the non-threatening turtle words and pictures. When in non-threatening scenario, the recognition accuracy of threatening pictures is significantly higher than that of the non-threatening turtle pictures. However, the recognition accuracy is not significantly different for threatening snake words and the non-threatening turtle words. The conclusions are as follows: (1) the memory recognition advantage exists in the vocabulary and pictures threatening information in the natural environment. And the memory of the pictures is better than that of the vocabulary; (2) memory advantage of threatening pictorial information still exists in the threatening and non-threatening scenario. When in non-threatening scenario, memory difference between the threatening and non-threatening vocabulary information is not significant.