Abstract: By making full use of the advantages of the advantages of the existing prediction model and selecting reasonably the prediction model, the dynamic workload of the network equipment will be predicted with the combination of the influential factors of the workload and the connection of the models in series or parallel to input or output the result. With the comprehensive utilization of the grey prediction and the BP neural network prediction model and the algorithm improvement of the prediction accuracy, this article analyzes and presents a comprehensive prediction model and gives the specific method and process of the prediction. The analysis of the result of the prediction has proved that this model has improved the prediction accuracy. This model and method work well in the quantitative prediction of the network equipment workload.