Abstract: Fire is an integral component of tropical savannas, but is often actively excluded from commercially grazed systems in northern Australia. The Kidman Springs Fire Trial (Victoria River District, NT), established in 1993, assessed the impact of fire management on woody cover and pasture condition. The trial is replicated on calcarosol and vertosol sites, with grazed experimental plots burnt early or late in the dry season, every two, four and six years, as well as unburnt controls. On calcarosols four-yearly late season fires were required to manage woody cover, whereas on vertosols, four-yearly early or late burns were adequate. Two-yearly fire or early dry season fire increased the proportion of dicots and suppressed the increase of perennial grass yield through time on the calcarosols; and on the vertosols, two-yearly or early fire increased annual grass yield but decreased total yield. Bio-economic modelling of a commercial cattle station found that implementing four-yearly fire improved animal production and enterprise profits, with late season fire providing the greatest benefits. There was an opportunity cost of implementing early season fire (as recommended for carbon and biodiversity outcomes) of $85/km2 compared to burning later in the year.