The 15N natural abundance (δ15N) technique offers good prospects for the quantification of the contribution of biological nitrogen fixation (BNF) to crop nutrition. The objectives of this study were to determine which plant parts best represented the whole plant with respect to δ15N and to quantify the BNF contribution to sugarcane. The experiment was carried out in pots of soil (100 kg pot−1) from the field. The sugarcane varieties used were RB867515 and RB92579 with the reference plants Brachiaria decumbens, millet (Pennisetum americanum) and sorghum (Sorghum bicolor). The experimental design was randomised blocks with four replications. The treatments were inoculation of the sugarcane with or without diazotrophic bacteria and the different parts of the sugarcane plant (root, green leaf, senescent leaf and different stem sections). The δ15N of different cane parts and varieties ranged from + 2.68 to + 5.85‰ after 365 days, while the δ15N of reference plants were from + 7.8 to + 8.3‰. The mean contribution of BNF to the sugarcane crop N supply was estimated to be 47% with no effect of variety or inoculation. The results indicate that the δ15N value of the entire shoot can be considered to represent the entire sugarcane plant, but no single organ could be relied on for this purpose.