Wildland fuelbeds are exceptionally complex, consisting of diverse particles of many sizes, types and shapes with abundances and properties that are highly variable in time and space. This complexity makes it difficult to accurately describe, classify, sample and map fuels for wildland fire research and management. As a result, many fire behaviour and effects software prediction systems use a generalised description of fuels to simplify data collection and entry into various computer programs. There are several major fuel description systems currently used in the United States, Canada and Australia, and this is a source of confusion for many in fire management. This paper (1) summarises the challenges of describing fuels, (2) contrasts approaches (association, classification and abstraction) for developing fuel description systems and (3) discusses possible future directions in wildland fuel description and science to transition to a universal fuel description system. Most discussion centres on surface fuel loadings as the primary descriptive characteristic. This synthesis paper is intended to provide background for understanding surface fuel classification and description systems and their use in simulating fire behaviour and effects, quantifying carbon inventories and evaluating site productivity.