The response surface model for dechlorination process of zinc oxide dusts by microwave was built with an analysis of variance (ANOVA). It was introduced to analyze the dechlorination experimental results, and the model fit well. The central composite design (CCD) was applied to optimize and investigate the influence of roasting temperature, holding time, stirring speed and those interactions to dechlorination efficiency. According to the prediction and optimization function of Design Expert software, the optimum conditions of dechlorination by microwave roasting were found. When controlling roasting temperature to 626 °C, holding time 25 min and stirring speed 54 rpm, the experimental result of dechlorination efficiency could reach 90.13% which is very close to the predicted dechlorination efficiency of 90.48%, indicating that the response surface methodology optimization of process parameters for dechlorination of zinc oxide dusts by microwave roasting are reliable.