Teleconnection is a linkage between two climate events that occur in widely separated regions of the globe on a monthly or longer timescale. In the past, statistical methods have been used to discover teleconnections. However, because of the overwhelming volume and high resolution of datasets acquired by modern data acquisition systems, these methods are not sufficient. In this paper, we propose a novel approach to finding teleconnections in global climate datasets using data mining technologies. We present experiments on real datasets and find some interesting teleconnections, including well-known ones such as ENSO. The experiments indicate that our method is usable and efficient.