Abstract Background and Aims The attractiveness of nectar rewards depends both on the quantity of nectar produced and on its chemical composition. It is known that nectar quantity and chemical composition can differ in plant species depending on the main pollinator associated with the species. The main aims of this study were to test formally whether nectar traits are adapted to pollination syndromes in the speciose Balsaminaceae and, if so, whether a combination of nectar traits mirrors pollination syndromes. Methods Comparative methods based on Ornstein–Uhlenbeck models were used to test whether nectar volume, nectar sucrose proportion, sugar and amino acid concentration and amino acid composition had evolved as a function of pollination syndromes in 57 species of Balsaminaceae. Cluster analysis and ordination were performed to derive clusters of species resembling each other in nectar composition. Key Results Evolutionary models for nectar volume and nectar sucrose proportion performed best when including information on pollination syndrome, while including such information improve model fit neither for sugar and amino acid concentration nor for amino acid composition. A significant relationship emerged between pollination syndrome and the combined nectar traits. Conclusions Our results show that nectar volume and nectar sucrose proportion evolve rapidly towards optimal values associated with different pollination syndromes. The detection of a signal indicating that nectar traits in combination are to a certain extent able to predict pollination syndromes in Balsaminaceae suggests that a holistic approach including the whole set of nectar traits helps us to better understand evolution of nectar composition in response to pollinators.