The distance information between nodes in P2P network is the basis for achieving topology-awareness which aims at optimizing the applications of overlay and solving the problems such as network monitoring. However, it seems infeasible to accurately and completely measure the distances between nodes due to the characteristics of P2P, such as being large-scale, self-organized, highly dynamic and so on. Consequently, researchers have put forward various prediction methods, and currently the network distance prediction technology is emerging as a new hotspot of research in P2P area. Firstly, a research framework is proposed, based on which the main aspects and the related technical issues of the research are analyzed. Meanwhile, the research history and the analysis of the classification are investigated. Many typical methods are introduced and compared. Lastly, the metrics of precision, as well as future research trends of network distance prediction is reviewed.