Fuel loads in forest areas are dependent on vegetation type and the time since the last fire. This paper reports a study on the feasibility of using remotely sensed data to estimate vegetative fuel loads. It describes two methods for estimating fuel loads using Landsat TM data based on equations describing litter accumulation and decomposition. The first method uses classification techniques to predict vegetation types coupled with fire history data to derive current fuel loads. The second method applies a canopy turnover rate to estimate litterfall and subsequently accumulated litter from biomass, thus utilising the dominant influence of canopy on remotely sensed data. Both methods are compared with data collected from Popran National Park in coastal New South Wales. The amounts of litter calculated with the biomass method were similar to field results, but the classification method was found to overestimate fuel loads. A sensitivity analysis investigated the impact of varying the vegetation constants and rates used in the fuel estimates to simulate uncertainty or error in their values. The biomass method was less subject to uncertainties and has potential for estimating fuel quantities to provide useful spatial information for fire managers.