In Japan, geothermal development has stagnated because of social risks, mainly the opposition of stakeholders. For mitigation of the risks, many studies have been conducted for analysis of social acceptance for development. Among them, Agent-Based Model (ABM) of opinion diffusion in a social network has been used to understand how acceptance develops dynamically. The objective of this study is to improve an existing model that reproduces realworld social network behavior through the introduction of a new parameter which reflects individual characteristics of stakeholders. We estimate the parameter by using a Bayesian network, introduce it into ABM and perform a proof-of-concept simulation of the model. This result gives confidence in the usefulness of this parameter as a practical tool for analysis of social acceptance in the future.