As part of the Cicerone Project's farmlet experiment, conducted on the Northern Tablelands of New South Wales, Australia, between July 2000 and December 2006, this study assessed the effects of varying soil fertility, pasture species and grazing management on the botanical composition of three 53-ha farmlets subjected to different management strategies. Starting with the same initial conditions, the farmlets were managed to reach different target levels of soil phosphorus (P) and sulfur (S); Farmlet A aimed at 60 mg/kg of Colwell P and 10 mg/kg S (KC₄₀) whereas Farmlets B and C both aimed at 20 and 6.5 mg/kg of P and S, respectively. Pastures were renovated on six out of eight paddocks on Farmlet A, but only one paddock of each of Farmlets B (typical management) and C (intensive rotational grazing) was renovated. Flexible rotational grazing was employed on Farmlets A and B (each of eight paddocks) while Farmlet C used intensive rotational grazing over its 17 major paddocks, which were further subdivided into 37 subpaddocks. This paper focuses on the botanical composition dynamics observed across all three farmlets and the explanatory variables associated with those changes. Eight assessments of botanical composition were carried out at approximately annual intervals across each of the 37 major paddocks distributed across the farmlets and the results for each of 49 species were aggregated into seven functional groups for analysis. The strongest correlation found was a negative curvilinear relationship between sown perennial grasses (SPG) and warm-season grasses (WSG). The most significant factors affecting the functional group changes were soil P, sowing phase, paddock and date. These factors led to significant increases in SPG and correspondingly lower levels of WSG on Farmlet A compared with Farmlet B. Farmlets B and C experienced similar, declining levels of SPG, and increasing levels of WSG suggesting that intensive rotational grazing did not lead to substantial changes in botanical composition, compared with flexible rotational grazing, in spite of the fact that intensive rotational grazing had much longer grazing rests and shorter graze periods than the other two farmlets. Soil P levels were also significantly associated with levels of cool-season annual grasses, legumes and herbs, especially on Farmlet A. In general, the largest differences in botanical composition were between Farmlet A and the other two farmlets; these differences were most closely associated with those plants categorised as sown, introduced, C₃pasture species. The levels of legume were generally low on all farmlets, due largely to the dry seasons experienced over most of the trial. Efforts to increase the legume composition on all farmlets were more successful on Farmlet A than on the other two farmlets due, presumably, to higher soil fertility on Farmlet A. Farmlet C, with its long rest periods and short graze periods, had a small proportion of legumes, due to the competitive effects of the accumulated tall grass herbage between grazings. The 'typical' management of Farmlet B also resulted in low levels of legume as well as increased 'patchiness' of the pastures and increased numbers of thistles.