Here, we describe the successful removal of Phenol Red from aqueous solutions by nanocobalt hydroxide. Also two approaches, artificial neural network (ANN) and Box–Behnken design (BBD), are used to investigate predictive models for simulation and optimization of the dye removal process. The effect of process variables (such as pH, sorbent dosage, and Phenol Red concentration) on the removal efficiency are investigated through performing the BBD. A training set for ANN is obtained using the same design. Statistical values show ANN model is superior to BBD model. Based on the validation data set, ANN model has higher value of coefficient of determination (R2: 0.99 ANN > 0.84 BBD), lower values of root mean square error of prediction (RMSEP: 3.17 ANN < 3.99 BBD), mean square error (MSE: 10.08 ANN < 15.95 BBD), and relative standard error of prediction (REP: 3.89 ANN < 4.89 BBD). According to obtained results, ANN has better prediction performance in comparison with BBD. In addition, the maximum removal efficiency (%Rmax = 99.9) is found at the initial concentration of dye = 68.08 (mg/L), pH = 2.0, and sorbent mass = 9.57 g/L by performing the BBD
