Despite a strong interest for understanding the link between attitudes, affect and performance at work, not many studies have investigated these relationships longitudinally. Longitudinal designs are highly relevant, because they allow inferring the direction of relationships between variables. Theoretical frameworks support the plausibility that both affective experiences influence performance and performance influences affective experiences, but this remains to be verified empirically. Therefore, focusing on the concept of psychological health at work, conceptualised as high well-being and low distress, this panel design study explores the relationship between psychological health and individual performance. A sample of teachers (n = 153) was used, with two measurement times separated by a one year lag. Results from hierarchical regressions show that within both well-being and distress, a specific component predicts performance: relationships with the social environment. Namely, social harmony (the social facet of well-being) and irritability-aggressivity (the social facet of distress) predicted individual performance one year later with similar effect sizes. In the reverse direction, also with a similar effect size, performance predicted two components of well-being: social harmony and serenity. This study suggests that the psychological health-performance relationship is bidirectional. Based on the results of this study, promoting good relationships between employees may be useful to enhance performance. On the other hand, offering a work context in which employees are able to perform might also improve their well-being in regards to themselves and others.