One of the main limiting aspects in the application of crown fire models at landscape scale has been the uncertainty derived to describe canopy fuel stratum. Available crown fuel and canopy bulk density are essential in order to simulate crown fire behaviour and are of potential use in the evaluation of silvicultural treatments. Currently, the more accurate approach to estimate these parameters is to develop allometric models from common stand inventory data. In this sense, maritime pine (Pinus pinaster Aiton) trees were destructively sampled in the South of the Iberian Peninsula, covering natural and artificial stands. Crown fine fuel was separated into size classes and allometric equations that estimate crown fuel load by biomass fractions were developed. Available crown fuel was determined according to the fuel load differences between un-burned and burned trees with similar characteristics. Taking our destructive post-fire inventory into account, available crown fuel was estimated as the sum of needles biomass, 87.63% of the twigs biomass and 62.79% of the fine branches biomass. In spite of the differences between natural and artificial stands, generic models explained 82% (needles biomass), 89% (crown fuel), 92% (available crown fuel) and 94% (canopy bulk density) of the observed variation. Inclusion of the fitted models in fire management decision-making can provide a decision support system for assessing the potential crown fire of different silvicultural alternatives. © 2014 Elsevier Ltd. All rights reserved.