In computational drug design, modelling studies are undertaken following two main strategies that depend on which information is available. If experimental data exist only for the molecules displaying the biological property of interest, a so-called ligand-based approach is taken; if information is available on the macromolecular target(s) of the compounds (e.g. proteins' 3D structures), target-based studies can be carried out. Recently, in the field of hERG K+-channel blocking drugs, pharmacophoric (ligand-based) studies started appearing aimed at determining the physicochemical features associated with the channel block, and also at predicting the hERG blocking potential of compounds. However, partial homology models (target-based) of the hERG channel have also been built and used as working tools to interpret electrophysiological and mutagenesis studies. Here, we review some of the ligand- and target-based in silico studies carried out on hERG, focusing on both their main characteristics and their meaning. In addition, we discuss some methodological aspects of the computational work that in our opinion should be considered, in view of the construction of reliable models possibly able to predict the functional behaviour of the channel system and the blocking potential of drugs.