Prediction of polypharmacology of known drugs and new molecules against selected multiple targets is highly useful for finding new therapeutic applications of existing drugs (drug repositioning) and for discovering multi-target drugs with improved therapeutic efficacies by collective regulations of primary therapeutic targets, compensatory signalling and drug resistance mechanisms. In this review, we describe recent progresses in exploration of in-silico methods for predicting polypharmacology of known drugs and new molecules by means of structure-based (molecular docking, binding- site structural similarity, receptor-based pharmacophore searching), expression-based (expression profile/signature similarity disease-drug and drug-drug networks), ligand-based (similarity searching, side-effect similarity, QSAR, machine learning), and fragment-based approaches that have shown promising potential in facilitating drug repositioning and the discovery of multi-target drugs.