Response surface method with central composite design was used to establish quadratic regression models and surface maps to relate panel properties, including static bending modulus of elasticity, modulus of rupture, internal bond strength, and thickness swelling with flake slenderness ratio, flake orientation, and panel density. A robot mat formation system was used to form the panels with predefined processing parameters. Results indicated that nonlinear models capable of including interactions were required to relate flake slenderness ratio, flake orientation, and panel density to panel properties, An optimization model was developed to obtain the best panel performance with respect to the three factors. The optimized combination of the three factors within the experimental range is: 133 for flake slenderness ratio, 8 degrees for surface flake orientation, and 0.62g/cm(3) for board density.