Abstract: The topology of social networks changes over time， which makes it very difficult to determine unknown spreading source. A localization method about diffusion source on dynamic networks is presented. The hidden source can be localized by means of modeling the network and deducing the spreading topology. First， the topological change of networks in the diffusion process is deduced based on the link analysis method given by hyperbolic geometry， and then the source based on the propagation topology is located. The large-scale experiments based on the actual networks and the synthetic networks show the feasibility of the proposed algorithm.

Key words: social network, dynamic evolution, information diffusion, source localization