In order to explore the impact mechanisms of different micro-break methods on cognitive load during work, a dual 2-back task-based experimental paradigm was used to investigate the effects of two rest conditions (phone use vs. eyes-closed rest) by collecting physiological signals and task performance data before and after breaks. The results show that different rest methods significantly influenced post-break performance, with both micro-break types reducing cognitive load levels. Post-break(electrodermal activity, EDA) differed significantly between groups, and although other physiological indicators did not reach statistical significance, the eyes-closed group exhibited a clearer downward trend in cognitive load. Physiological signals demonstrated higher sensitivity than performance metrics in capturing implicit cognitive resource allocation. It is concluded that compared to phone use, breaks that avoid consuming additional cognitive resources (e. g., eyes-closed rest) are more effective in optimizing workplace rest strategies, thereby enhancing work efficiency. The combined use of performance and physiological measures provides a more comprehensive and accurate quantification of cognitive load assessment.