Within the context of early drug discovery, a new pharmacophore-based tool to score and align small molecules (Pharao) is described. The tool is built on the idea to model pharmacophoric features by Gaussian 3D volumes instead of the more common point or sphere representations. The smooth nature of these continuous functions has a beneficent effect on the optimization problem introduced during alignment. The usefulness of Pharao is illustrated by means of three examples: a virtual screening of trypsin-binding ligands, a virtual screening of phosphodiesterase 5-binding ligands, and an investigation of the biological relevance of an unsupervised clustering of small ligands based on Pharao.