Background: The relationship between positivity, i.e., the proportion of positive and negative emotions, and job burnout may be of a curvilinear shape. From a theoretical point of view, it is a causal relationship, since positivity can be regarded as a proximal dimension of occupational well-being, and burnout as a distal one. However, previous studies have been mostly cross-sectional and have tested only linear relationships between these variables. Therefore, the aim of the study is to examine the shape of the relationship between positivity and burnout using both cross-sectional and longitudinal designs, on the example of civil servants. Material and Methods: The study involved 238 civil servants (73.5% of whom were women). Positivity was assessed with the Job-related Affective Well-being Scale. Exhaustion and disengagement, 2 components of job burnout, were measured twice, at a 4-month interval, using the Oldenburg Burnout Inventory. Results: The cross-sectional models assuming the curvilinear relationship between positivity and job burnout were better fitted to the data than models with the linear relationship only. Thus, positivity was curvilinearly related to both exhaustion and disengagement, with an inflection point around 2. In the longitudinal models, for exhaustion the curvilinear model was again better fitted to the data, while for disengagement it was the linear model. Conclusions: The relationship between positivity and exhaustion was curvilinear among the civil servants. This may indicate personal costs of maintaining a high ratio of positive to negative emotions at work. Nevertheless, the role of positivity for disengagement seems more complex, with a possible long-term protective function. Med Pr. 2019;70(2):201–11