Abstract: Based on the data of audience rating, a model was developed to predict the number of audience who will respond to an advertisement, thus solving the problem that there are no effective indices to evaluate the efficiency of advertising activities. The model is composed of an exposure frequency distribution submodel and a prediction submodel based on the former one, and both are described for the estimate of their parameters. The genetic algorithm method is introduced to estimate the parameters of the prediction submodel. With the real data of a financial organization taken as example to verify the model developed, comparative results between predicted and actual number of the audience responding to an advertisement showed that the proposed model can offer a relative accurate prediction without the support of deciders' experience if the data of audience rating are easy to acquire. Therefore, the proposed model can be used to provide quantitative evaluation indices in practice at advertising activities.