Large Language Models (LLMs) have shown impressive capabilities across various fields, including across social sciences. While research has explored LLMs’ emotion capabilities, studies on coping mechanisms are limited. This work explored LLMs’ ability to represent emotion-focused coping, specifically updating beliefs and goals in response to emotional events. Coping has broad influences on physiological, cognitive, and social processes. We conducted two experiments on leading LLMs: a free-choice paradigm and a hurricane scenario. In the first experiment, LLMs did not exhibit preference changes between easy and hard choices, unlike humans. In the second experiment, LLMs showed no difference in hurricane perceptions based on evacuation decisions. Our findings suggest that current LLMs do not change or predict changes in beliefs and goals similarly to humans when coping with emotions, highlighting their limitations in modeling and predicting coping behaviors.