According to the design principle of the central composite experimental, the method of response surface analysis with three factors and three levels was adopted based on one factor test. A second-order quadratic equation for photocatalysis of Procion Red MX-5B was built. Response surface and contour were graphed with the decoloration rate of Procion Red MX-5B as the response value. Based on the analysis of the response surface plots and their corresponding contour plots, effects of pH value, irradiation time and catalyst loading were explored. By using this new method, the optimum decoloration condition was obtained as follows; pH value, 1. 3; irradiation time, 49.9 min; catalyst loading, 0.57 g/L. In the optimization, R-Squared and Adj R-Squared correlation coefficients for quadratic model were evaluated quite satisfactorily as 0. 9310 and 0. 8620, respectively. Under the optimum conditions established, the performance of 99.47% for color removal was experimentally reached. It was found that all factors considered have an important effect on the decolorization efficiency of Procion Red MX-5B. By the ANOVA analysis and model confirmation the optimal solution obtained using RSM was experimentally validated and credible with preferable instructional ability for experiments.