This study describes a lumped kinetic modelling architecture with which to simulate and predict the pyrolysis reaction kinetics of hemicellulose biopolymers. A key feature of the model is the replication of essential compositional features of biomass hemicelluloses into the chemical reaction pathways of the model. More specifically, five new model compounds are proposed to reflect the contribution of acetylated glucuronoxylan, arabinoxylan, (galacto) glucomannan, xyloglucans and beta-glucan toward replicating the pyrolytic reactivity of hemicellulose. Following a careful review of the literature, available experimental data of the pyrolytic conversion of pure hemicellulose polysaccharides, studied as a function of heating rate in the ranges 0.16-60,000 K s(-1), is used to train the kinetic model. The reactivity and species evolutions of pentose- and hexose-based polysaccharides is shown to be distinctive. This reaction kinetic behaviour is described in the model by the definition of separate model components for each biopolymer. The pyrolytic reactivity of each polymer is prescribed to the model by optimising Arrhenius and reaction pathway descriptors to a series of data sets harvested from the literature. The so produced model is shown to have improved accuracy in predicting the temperature, and composition-dependent pyrolytic reactivities of hemicellulose biopolymers. (C) 2017 Elsevier B.V. All rights reserved.