► Optimization of color removal and COD removal done by response surface approach. ► The experiments were designed using Box–Behnken spherical design. ► Two quadratic polynomial models were developed for the responses. ► Single point numerical optimization was done considering three constraints. ► Validation by performing the experiment under optimized conditions. Photocatalytic degradation of methyl blue (MYB) was studied using Ag+ doped TiO2 under UV irradiation in a batch reactor. Catalytic dose, initial concentration of dye and pH of the reaction mixture were found to influence the degradation process most. The degradation was found to be effective in the range catalytic dose (0.5–1.5g/L), initial dye concentration (25–100ppm) and pH of reaction mixture (5–9). Using the three factors three levels Box–Behnken design of experiment technique 15 sets of experiments were designed considering the effective ranges of the influential parameters. The results of the experiments were fitted to two quadratic polynomial models developed using response surface methodology (RSM), representing functional relationship between the decolorization and mineralization of MYB and the experimental parameters. Design Expert software version 8.0.6.1 was used to optimize the effects of the experimental parameters on the responses. The optimum values of the parameters were dose of Ag+ doped TiO2 0.99g/L, initial concentration of MYB 57.68ppm and pH of reaction mixture 7.76. Under the optimal condition the predicted decolorization and mineralization rate of MYB were 95.97% and 80.33%, respectively. Regression analysis with R2 values >0.99 showed goodness of fit of the experimental results with predicted values.