Artificial neural network (ANN) technique is applied to make forecasting of summer rainfall patterns for east China in the following way: i) predictors are searched in the previous datases following the relation of the pattern to the circulation and SST features of the preceding winter; ii) ANN prediction is constructed for the pattern in terms of the anterior 40 year data as a training sample for the ANN to learn according to specified rules, thereby arriving at a forecasting model of a categorized pattern. Also, attempt was made to predict such patterns in 1992～1996, with the results rather close to the observed.