Hydrogen is gaining advantages in combustion performance and zero-carbon emissions and widely regarded as a key element of an energy solution for the 21^st century. The hydrogen evolution reaction (HER) method, a method of producing hydrogen, is a very important tool for producing green hydrogen that does not produce CO₂. However, searching for HER catalysts for high hydrogen production costs a lot because they involve a lot of experiments in experimental methods and a lot of computations in search spaces in computational methods. To overcome this drawback, we intend to use machine learning in HER catalyst research. HER catalyst research using machine learning not only reduces computational costs but also reduces the number of experiments. The study of catalysts using machine learning initially requires similar costs to general experiments and calculations for data collection, but machines that have been trained will then be able to search for catalysts at lower costs. The learning of the Machine was conducted using molecular structure and information on atoms, and energy for hydrogen adsorption was well predicted through the input information. We anticipate that this study of catalysts using this machine learning is expected to contribute to the discovery of new catalysts and the expansion to other reaction catalyst research.