Plenty of SO2 is generated in the productive process of sulfuric acid in chemical fertilizer industry. In order to improve the efficiency and accuracy of SO2 emission prediction, a new method combining extreme learning machine (ELM) and genetic algorithm (GA) for SO2 concentration prediction was proposed based on analyzing the influencing factors of sulfuric acid productive process. In order to avoid the influence on predicting effect of ELM by the randomness of input weight matrix and hidden layer deviation, GA was used to optimize the input weight matrix and hidden layer deviation, and the GA-ELM model for SO2 emission prediction was built. Case analysis was made using statistical data of sulfuric acid productive process in a chemical fertilizer factory. The result shows that the prediction on possibility of SO2 emission by GA-ELM model can be relatively accurate and effective.