This study is an application of an experimental design methodology for the optimization of the culture conditions of alpha-l-arabinofuranosidase production by Talaromyces thermophilus. Wheat bran and yeast extract were first selected as the best carbon and nitrogen sources, respectively, for enzyme production. A Plackett-Burman design was then used to evaluate the effects of eight variables. Statistical analyses showed that while pH had a negative effect on alpha-l-arabinofuranosidase production, wheat bran and MgSO4 had a significantly positive effect. The values of the latter three parameters were further optimised using a central composite design and a response surface methodology. The experimental results were fitted to a second-order polynomial model that yielded a determination coefficient of R (2) = 0.91. The statistical output showed that the linear and quadric terms of the three variables had significant effects. Using optimal conditions, the experimental value of alpha-l-arabinofuranosidase activity produced was very close to the model-predicted value. The optimal temperature and pH of enzyme activity were 55 A degrees C and 7.0, respectively. This enzyme was very stable over a considerable pH range from 4 to 9. The crude enzyme of T. thermophilus rich in alpha-l-arabinofuranosidase was also used for saccharification of lignocellulosic materials and arabinose production.