Jackbean is an annual legume frequently used as green manure in tropical intercropping systems with bananas. Although the beneficial effect of nitrogen (N) release from above-ground residues on banana nutrition is well known, little information is available on the N transfer from jackbean roots before and after the above-ground harvest. The aim of this study was to assess the relative contribution of exudates and root turnover in the N transfer from jackbean to banana in a greenhouse experiment. Nitrogen transfer was studied in a 6-month trial using the 15N leaf feeding method, and estimated with a box model of 15N dilution based on the observed data of 15N content in exudates and decomposing roots. For the sowing–harvest period, the amount of N transferred from jackbean exudates represented 16% of banana N uptake and 0.7% of jackbean N uptake. Therefore, the N transfer flux via exudates was 23 times higher in terms of N input for the recipient plant than in terms of N output for the donor plant. This value, which is an index of the effectiveness of N transfer, was lower than those reported previously for other soil–plant systems in greenhouse conditions. This would be due to differences in root traits of the recipient plants. The amount of transferred N from root turnover after jackbean harvest represented 52% of banana N uptake in that period. The box model described N transfer from both legume N sources adequately (r2 = 0.92). For the whole experiment, 38% of banana N uptake was derived from jackbean (6% from exudates and 32% from root turnover), and 62% from soil N. The results indicated that N transfer from root exudates of jackbean would be a useful but minor process compared with N release from root turnover in soil. The experimental and theoretical approach proposed in this study may be useful in screening studies to assess the capability of herbaceous legumes to transfer N.