Over the last few decades, there are steady development in the understanding of electrocatalytic reactions on the heterogeneous catalysts for polymer electrolyte membrane fuel cells (PEMFC). Unfortunately, high performance catalysts for electrochemical systems are made by a huge amount of experimental trial and error process. First-principles density functional theory (DFT) calculations have begun to provide a rapidly screened database and a simple way to rationally design materials. These atomic-scale database are suitable for guiding of heterogeneous catalysts design. Despite this, high-throughput screening cannot adequately handle the kinetics such as reaction barrier. Database for transition state energies demands considerable time and computational costs, therefore the other concept is required. In this work, we attempt to mitigate time and computational expenses by using a relationship between the reaction energy and the activation barrier. Also, machine learning techniques are applied to predict the key parameter of the dissiciation reaction barrier.