Data from the Cicerone farmlet study were used to quantify the balance between pasture feed supply and the demand from grazing livestock, in terms of metabolisable energy (ME), on three differently managed farmlets (each of 53 ha) on the Northern Tablelands of New South Wales, Australia. Farmlet A had a high level of pasture renovation and higher soil fertility than the other two farmlets and employed flexible grazing management over eight paddocks. Farmlet B was designed to represent management 'typical' of the region and had the same grazing management and number of paddocks as farmlet A but moderate levels of pasture renovation and soil fertility. The third farmlet (C) had the same level of inputs as farmlet B but practised intensive rotational grazing over 37 paddocks. Regular measurements of the feed supply, namely herbage mass and quality, pasture growth and supplement fed and of feed demand were assembled to provide monthly estimates of the balance between feed supply and animal demand of all classes of livestock run on the experiment over its duration of 6.5 years. The significantly greater stocking rate, liveweight and reproductive rate of sheep reached on the higher input system (farmlet A) meant higher levels of ME were required to satisfy the nutritional demands of these animals. As only limited measurements were taken of animal intake, it was assumed that the supply of ME was derived from pasture growth and supplement fed. Using key livestock management dates and measurements of liveweights, the changes in the energy requirements of each class of animal were calculated and aggregated to provide an estimate of overall livestock energy demand over time. Subtracting the energy demand from the estimated energy supply provided a partial net energy balance. Measurements of the rates of change of green herbage during grazing events were found to be highly dependent on stock density with farmlets A, B and C recording rates of change of up to -50, -30 and -200 green DM/ha.day, respectively. Over a series of generally drier-than-average years, the ME supplied in pasture growth and through supplementation was at times inadequate to meet the energy demands of the livestock, resulting in periods during winter when the partial energy balance became negative. Similar feed deficits were observed for all three farmlets, suggesting that they were overstocked to a similar extent. In spite of the divergence in the stocking rate supported by each farmlet, the similarity of the ME balances between farmlets suggests that no farmlet was subjected to bias because of decisions relating to feed supply and demand. The analyses presented suggest there is considerable potential for practical paddock and grazing management to be improved if more timely and regular assessments can be made of changes in the feed energy supply using satellite images of normalised difference vegetation indices and feed energy demand using calculations of the ME required by grazing livestock.