The high heterogeneity of heating, ventilation, and air conditioning (HVAC) systems and smart building applications makes it difficult to improve decarbonization via horizontal deployment. This study investigates a mechanical evaluation method for the similarity of HVAC systems and a modeling policy for the evaluation, which will serve as a matching indicator for the horizontal deployment of applications. In this study, 11 datasets that describe the heat source system as a resource description framework were created, and an algorithm for extracting the maximum common substructure (MCS) of a chemical compound was applied. Specifically, using RDKit, a cheminformatics library, the same construction increase was extracted and visualized from the dataset that was converted to a description system for chemical compounds called SMILES. Subsequently, a list of points and problems that could be applied to the similarity evaluation of HVAC systems was created. As a result, it was confirmed that RDKit can accelerate the extraction and visualization of similar structures. However, because the connection direction between elements is not considered in chemical compounds, in certain cases, the similarity of systems with different connection orders has been highly evaluated. In conclusion, the findings provide many insights into the digitization method for chemical structures and the MCS algorithm that can be applied to HVAC systems. By contrast, grouping equipment with the same role for redundancy and dealing with the connection of equipment in a directed graph remain problems.