There is ample empirical evidence for negative effects of emotional labor (surface acting and deep acting) on workers' well-being. This study analyzed to what extent workers' ability to recognize others' emotions may buffer these effects. In a 4-week study with 85 nurses and police officers, emotion recognition moderated the relationship between emotional labor and work engagement: Workers with high emotion recognition engaging in emotional labor did not report lower work engagement after 4 weeks, whereas those with low emotion recognition did. These effects pertained to both surface and deep acting. The results suggest that emotional labor be not necessarily detrimental to workers' engagement. Instead, the impact of emotional labor hinges upon workers' ability to correctly identify interaction partners' emotions. (Contains 2 tables and 2 figures.)