We describe developments to the land surface model JULES, allowing for flexible user-prescribed harvest regimes of various perennial bioenergy crops or natural vegetation types. Our aim is to integrate the most useful aspects of dedicated bioenergy models into dynamic global vegetation models, in order that assessment of bioenergy options can benefit from state-of-the-art Earth system modelling. A new plant functional type (PFT) representing Miscanthus is also presented. The Miscanthus PFT fits well with growth parameters observed at a site in Lincolnshire, UK; however, global observed yields of Miscanthus are far more variable than is captured by the model, primarily owing to the model's lack of representation of crop age and establishment time. Global expansion of bioenergy crop areas under a 2 ?C emissions scenario and balanced greenhouse gas mitigation strategy from the IMAGE integrated assessment model (RCP2.6- SSP2) achieves a mean yield of 4.3 billion tonnes of dry matter per year over 2040-2099, around 30 % higher than the biomass availability projected by IMAGE. In addition to perennial grasses, JULES-BE can also be used to represent short-rotation coppicing, residue harvesting from cropland or forestry and rotation forestry.