Given the growing demand for studies dealing with natural disasters, the research fields of emotion and social cognition require validated picture stimuli of natural hazards. Such material is essential for studying perceptual processes and behaviors of exposed individuals, and it could find practical applications, such as the improvement of communication strategies during crises. We present the Natural Disasters Picture System (NDPS), a database of pictures of natural hazards, with an emphasis on volcanic threats, and their impact on the environment and humans. We first describe in detail the picture selection and database creation. We then report the validation procedure. One hundred twenty participants rated the pictures on the basis of four dimensions: valence, arousal, dominance and certainty. For each picture, we ultimately determined the best-fitting emotion on the basis of its dimensional pattern. The Hierarchical Ascendant Classification, which yielded 4 clusters subdivided into 9 classes, indicated a highly consistent and distinctive classification of the pictures. Overall, 90% of the pictures elicited negative emotions (fear or sadness), and the other 10% induced neutral to positive emotions (e.g., aesthetic emotions). The NDPS offers a new tool for studying natural events and disasters in the field of affective and cognitive sciences, which will benefit from scientific research and its practical applications. The NDPS is unrestrictedly accessible for researchers.