The amount of reactive nitrogen has more than doubled in terrestrial ecosystems due to human activities such fertiliser application that is predicted to increase dramatically in coming decades. We conducted a 3-year experiment in a Neotropical savanna in which we determined the effects of increased N deposition on litter decomposition in plots subjected to different levels of N addition (50 kg N ha−1 year−2, 20 kg N ha−1 year−2, or no N addition). For this, we compared the litter decomposition from the bunchgrass Tristachya leiostachya using litter collected from plots with different N addition treatments. Five randomly selected bags of litter from each N addition treatment (origin) were distributed to each plot (destination). We also compared litter nitrogen (N) concentration and indicators of microbial activity (basal respiration, carbon of microbial biomass, metabolic quotient, enzyme activity) in all plots. We found that nitrogen addition influences litter decay, but in idiosyncratic ways that differ between years. In year 1, litter decomposed faster in high-N addition plots than in low-N and control plots, regardless of its origin. In contrast, litter from high-N addition plots decomposing fastest in year 2, regardless of its destination. Finally, there was no effect of either litter origin or destination on the rate of decomposition in year 3. Litter collected in high-N addition plots had a concentration of N 12–17% higher than litter collected in other plots and higher in 2009 than in other years. Four years after the beginning of the fertilisation experiment, concentration and the microbial activity in the soil did not differ between the treatments. Our findings suggest that the levels of N addition predicted for Neotropical savannas can alter litter N concentrations and the process of litter decomposition, but that the direction and magnitude of these changes may be challenging to predict since that precipitation can influence the mechanisms regulating decomposition in the Cerrado.