This experiment was a 3 x 3 x 2 factorial laid out in a randomized complete block design and was replicated three times. Treatments comprised three sward types (pure grass, pure legume and grass/legume mixed swards), three cutting regimes (4 weeks, 8 weeks and 12 weeks) and two nitrogen fertilizer rates (0 and 300 kg N ha-1). Herbage production was 16% and 69% higher (P < 0.05) on grass-legume swards in 2006 and 2007, respectively, compared with production from pure grass swards. Grass dry matter yield as a proportion of the total herbage dry matter was significantly (P < 0.05) increased from 45% to 61% in 2007 with 12-weekly interval compared with 4-weekly interval. The weed proportion was significantly reduced from 28% to 19% with the infrequent cutting interval of 8 weeks compared with the 4 weeks in 2007. Cutting frequency treatment did not influence the proportion of legume in 2006 and 2007. However, the legume dry matter yield was generally increased with longer intervals between cuts. Fertilizer-N application had no effect on the proportions of grass, legume and weed in 2006 and 2007.