Inclusion of forage legumes in low-input forage mixtures improves herbage production and soil fertility through addition of nitrogen (N) from N 2-fixation. The impact of different grass–legume mixtures on the N contribution of the forage mixture has rarely been investigated under comparable soil and climatic conditions. We conducted a field experiment on a sandy soil at two nitrogen levels with seven two-species forage mixtures: alfalfa, bird's-foot trefoil, red clover, or white clover in mixture with perennial ryegrass, and white clover in mixture with meadow fescue, timothy, or hybrid ryegrass. We found high N 2-fixation of more than 300 kg N ha −1 from both red clover and alfalfa even when the two mixtures received 300 kg total-N ha −1 in cattle slurry. The addition of cattle slurry N fertilizer lowered N 2-fixation for white clover and red clover as expected, but for bird's-foot trefoil and alfalfa no changes in the proportion of N derived from N 2-fixation was observed. We conclude that the competition for available soil N from perennial ryegrass in mixture was an important factor for the proportion of N in alfalfa, white clover, and bird's-foot trefoil obtained from N 2-fixation. White clover had a high proportion of N derived from atmosphere for all companion grasses despite significant differences in white clover proportion. Although the perennial ryegrass–alfalfa mixture in the grass phase yielded more than twice the N from N 2-fixation compared to white clover in the perennial ryegrass mixture, this did not in the following year lead to higher residual N effects of alfalfa. Both in terms of N yield in the grass phase and N yield in the subsequent spring barley red clover contributed most to the improvement of soil N fertility.