Recently, the prospect of developing molecular devices on solid surfaces requires treating molecular clusters or aggregates adsorbed on these surfaces. However, in contrast to progress in experiment, using computer simulation for this subject has barely been developed yet. The reason is that ab initio calculations, which provide accurate results, are computationally too demanding even for cluster models, and the reliability of semi-empirical calculations, which can reduce computational costs, is questionable for interactions between adsorbed molecules and ones between adsorbates and the surface. These circumstances require developing a calculation method that not only reduces computational costs for treating species adsorbed on a solid surface, but also provides moderate accuracy. Currently we are developing[1] a quantum mechanical/molecular mechanical (QM/MM) method to treat species adsorbed on noble metal surfaces that reduces computational costs. The QM/MM method describes the interaction between adsorbates and the surface in a molecular-mechanical way, whereas the adsorbates are treated quantum- mechanically. In our QM/MM method, analytical atom-atom potential functions, i.e., Lennard-Jones-type functions, are used for repulsion and dispersion interactions between the adsorbates and the surface. The adsorbate-surface- induction interaction is also included in the calculations by assuming that image charges appear in the metal. Although our method still lacks accuracy in calculating the interaction between the adsorbates and the surface, it describes the interaction between the adsorbed molecules accurately. Thus it should be a promising method for analyzing the self-assembly process on the surface in physisorption, in which the distance between the surface and the adsorbate is large and the adsorbate-surface interaction is small. In fact, the application of the developed QM/MM method for the examination of several molecules adsorbed on noble metal surfaces[2] showed that the calculated results by our method are consistent with experimental observations.