There is growing evidence for a positive non-linear response in N2O emissions where increasing N fertilizer rates critically exceed crop requirements from a range of cropping systems. However, the majority of these studies have been conducted in temperate regions, and it remains unclear if the non-linear response hypothesis can be transferred to tropical regions due to their inherently lower carbon (C) content and fertility. This study assessed the impact of different nitrogen (N) fertilizer application rates in combination with a carbon source (green-waste compost) on soil nitrous oxide (N2O) and carbon dioxide (CO2) emissions and crop productivity from a maize-soybean rotation in tropical Sri Lanka. Low cation exchange capacity and extremely high rainfall promoted higher annual cumulative N2O emissions than reported from other cropping systems, ranging from 2.4–9.16 kg N2O-N ha−1 yr−1 for the different fertilizer rates (0 N, 100 N, 220 N, 300 N). The ΔN2O emitted for every additional kilogram of fertilizer N applied (i.e. the marginal N2O emission), declined from ∼24 g kg−1 for the first 0−100 kg N applied to 7 g kg−1 at 300 kg N. This resulted in declining Emission Factors with increasing N rate, as factors other than N limited denitrification. The addition of compost removed this limitation, resulting in constant ΔN2O production rates and Emission Factors with increasing N inputs, suggesting low C availability limited N2O production at high N rates. The study shows that the non-linear response hypothesis might not hold for highly weathered tropical soils, where C availability limits N2O emissions at high N inputs. While the use of compost can be a viable option to increase soil fertility in these highly weathered tropical soils, it needs to be balanced with reduced N fertilizer application rates to avoid elevated losses of N2O.