A ligand-based 3D-QSAR study for the identification of MMP3 inhibitors was developed by applying an innovative alignment method capable of taking into account information obtained from available X-ray MMP3 structures. Comparison of the obtained model with data recently published using a docking-based alignment method indicated that the ligand-based 3D-QSAR model provided better predictive ability. A second external test set of 106 MMP3 inhibitors further confirmed the predictive ability of the 3D-QSAR model. Finally, certain iminodiacetyl-based hydroxamate-benzenesulfonamide conjugates, which were predicted to be active by the 3D-QSAR model, were tested in vitro for MMP3 inhibition; some provided low nanomolar activity. As such, our results suggest that the multitemplate alignment method is capable of improving the quality of 3D-QSAR models and therefore could be applied to the study of other systems. Furthermore, since MMP3 is an important target toward the treatment of arthritis, this model could be applied to the design of new active MMP3 inhibitors.