In previous work, young children avoided associating with a wrongdoer, despite incurring a personal cost. Such aversionto wrongdoers, arguably a reflection of moral development, weakens when the cost becomes very large (Tasimi & Wynn,2016). We model this moral decision-making process using the nave utility calculus (Jara-Ettinger et al., 2016), assumingutility maximization amidst uncertainty using Bayesian framework. The cost is defined as the number of stickers forgoneby choosing a nice persons smaller offer over a mean persons larger one, following the ratios of 1:2, 1:4, 1:8, and 1:16. Ourmodel aims to explain previous findings, and test predictions for new ratios. Compared to a baseline condition where nobackground information is available, children are predicted to choose the nice person when the cost is low, but reverse theirpreference when the cost becomes increasingly high, which would suggest a utility account for moral decision making.