In this paper, a new model system is introduced, which synthetically applies time series model, nonlinear regression and combination forecasting model to forecast the change of the market price.

Finally, the results show the methods can effectively come into being regression analysis model of time-series data streams, and fulfill the prediction of future data streams.

The forewarning control limits and time series regression judgment model are built up based on the reliability series model in a non-repairable system and Shewhart control charts.