response surface methodology (RSM) and artificial neural networks (ANN) towards efficient optimization of flexural properties of gypsum-bonded fiberboards. CERNE, v. 24, n. 1, p. 35-47, 2018. HIGHLIGHTS The higher non-wood extractives causes to the higher setting time of the gypsum paste, while temperature decreases. The ANN prediction model is a quite effective tool for modeling bending strength of gypsum-bonded fiberboard. Maximum MOR is achieved by increase in bagasse, kenaf and glass fibers content and reaches to 10.81 MPa and 11MPa by RSM and ANN at optimum condition.