AbstractIn the first stage of pulp bleaching, the quantity of added chemicals (ClO2 and/or Cl2) is commonly controlled by kappa factor, based on a kappa number online analyzer together with a compensated brightness control scheme as a feedback strategy. However, a kappa number analyzer is not always available, so the bleaching quality relies heavily on the chemical dosage set-point chosen by the operators. In this study, an improved model for the chlorination stage brightness optimization was proposed, based on brightness and residual chemicals before pulp enters the bleaching tower. Additionally, the experience of operators of (C95/D5) bleaching was employed in order to find an optimum chemical dosage set-point quickly. The golden section search algorithm (i.e., ‘0.618 method’) was used to find the optimum chemical dosage in this paper. After applying the proposed method in a pulp mill (C95/D5) bleaching stage, the chlorination stage brightness shifted from 62.9% ISO to the target value 60.7% ISO. Meanwhile, the standard deviation was reduced from 3.0 to 2.5.