Abstract: Taking the concurrent project multi-period steel procurement problem as the research project, this paper establishes steel price time series prediction model through the stationarity test, pure random inspection and seasonal analysis processing by selecting the increasing steel historical price data in a certain period as sample. Then combining steel ordering cost with financing cost in an order model, the paper finally gets optimal strategy for concurrent project multi-period steel procurement strategy. Research results show that the prediction model and purchasing model can solve concurrent project multi-period steel procurement and inventory problems under the condition of increasing prices, and compared with other strategies, the corresponding purchasing storage strategy can save cost under the condition of different financing rates.

Key words: supply chain management； price forecasting； time series； concurrent project, multi-period steel procurement