Some alternative, unconventional plant raw materials can provide large amounts of biomass for producing quality cellulose pulp. One such raw material can be Hesperaloe funifera plants, that grows with little water requirements, and which fibre morphology is especially suitable for making cellulose pulp. It uses a factorial design of experiments applied to pulping with diethanolamine and apply polynomial and neural fuzzy models in order to predict the properties of pulp (contents of holocellulose, alpha-cellulose and lignin, yield, viscosity, Kappa number and beating grade) and paper sheets (tensile index, stretch, burst index, tear index and brightness) according to the operating variables: temperature (155-185 degrees C), time (30-90 min) and diethanolamine concentration (50-80%). The polynomial and neural fuzzy models tested reproduce the experimental results with errors less than 8-10% for all dependent variables, with the exception of tear index which only reproduces 50% of results with errors less than 10%. (C) 2010 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.