With the arrival of the information age, people are faced with a large number of information resources on Internet, in order to solve the problem of information overload, the recommendation algorithm has been used in lots of information systems and Internet applications, however, most traditional recommendation systems have problems with cold start and recommended homogeneity. This paper introduces the relevant theories of chance discovery, discusses the advantages of chance discovery, which can connect the weak signals demand with the implicit related resources, and proposes a personalized recommendation algorithm based on the chance discovery so as to dig deeper into the potential requirements and preferences of users. In the experiment, we not only consider the precision but also refer to the diversity and novelty as the evaluation index. Through the extensive experiments, we compared with traditional recommendation algorithms, and then it is proved that our algorithm is helpful to improve the quality of the recommendation.