Children understand the world through learning. But they cannot obtain knowledge only through direct observation. Such as ethics, morality, and religious beliefs, children need to obtain knowledge from others. But children are likely to be deceived or gain wrong knowledge if they trusted the information indiscriminately. Therefore, as one of the basic abilities required for children's learning, the selective trust could help children gain useful information and construct their cognitive world. Selective trust germinated in infancy and matured from toddler to childhood. Current research results showed that children's selective trust was a credulous and rational process. So, what was the true nature of children's selective learning? How to explain these inconsistent findings? What was the cognitive mechanism behind these inconsistencies? The dual-processing model of selective trust provided a new way for explaining the cognitive mechanism behind children's learning. A core point of the dual-processing model was that children use simple heuristic strategies to make default judgments in the early stages of development. However, as they grew up and their cognitive abilities developed, children's trust judgment strategies became more rational. The heuristic strategy was the Type I processing, and the rational processing strategy was the Type II processing. When Type II processing appeared, Type I processing would not disappear but still coexist with Type II processing in children's selective trust. When the Type I processing could not work, Type II processing would intervene and help children make rational choices. Unlike the dual-processing model in cognitive psychology, the character of the dual-processing model in selective trust was that, in selective trust, Type I processing was based on simple heuristic strategies and did not require specific ability and knowledge. Previous research raised that Type I process included at least two main strategies: general-trust heuristic and trust-the-better heuristic. We proposed a new strategy, associative heuristic. Under this heuristic, children only need to make trust judgments by associating surface features with information provider characteristics based on their observation of the information provider's previous behavior. Type II processing required children to use rational processing methods to compare and match the information providers' abilities and traits more accurately according to specific task requirements. Type II processing had two manifestations. The first aspect was that children could trust appropriate information providers based on the clues and specific task scenarios given when faced with a single trust clue. The second aspect was that children could overcome the interference of social cues and make trust judgments based on cognitive cues when faced with multiple trust cues. After reviewing the relevant studies, we proposed that the cognitive ability, the family environment, and the nature of the task would affect children how use the two processing strategies. In sum, based on introducing the dual-processing model, this article sorted out the credulity and rationality of children under different selective trust task situations, and discussed how to distinguish between Type I and Type II processing and what factors affect them. Discussing this model could help researchers understand the relevant results in children's selective trust and clarify the mechanism behind it, which can help researchers to better understand children's learning process.