Model-based inferences are used commonly in psychological research. Consequently, model comparison has become increasingly important as a scientific procedure. However, while marginal likelihood is a fundamental quantity for model comparison, it is difficult to calculate. The widely applicable Bayesian information criterion (WBIC) is used for calculating marginal likelihood in various situations, but its calculation procedure is not well known in the psychology field. Accordingly, this study introduces WBIC using the Just Another Gibbs Sampler (JAGS) language, which can easily write a Markov chain Monte Carlo estimation procedure in Bayesian data analysis. Further, it provides two real data examples to demonstrate ways to define the WBIC’s weighted likelihood. The JAGS codes are opened and used for analysis.