Research and development of novel drugs cost too much time and money at high risk. Drug repositioning, which is to predict different therapeutic indications for approved drugs based on various technologies would help to reduce time, costs and risks of drug development. Experimental approaches alone are not sufficient to find new indications for approved drugs due to the huge amount of diseases and existing drugs. The integration of theoretical designs and experimental approaches along with published data of omics and drug informatics could lead to a new stage of drug development. Theoretical prediction of drug repositioning has provided a crucial direction to the research community of computational biology and systems biology. The current strategies of drug repositioning based on computing technology are underlined in this paper, namely drug-target relationship, drug-drug relationship and drug-disease relationship. Here we review the reported technologies and methods in this field with success cases at present.