Abstract: This paper proposes an optimal aggregation point selection(OAPS) algorithm based on cost of data collection for industrial wireless networks,which regards aggregation point selection as a location problem in graph theory.The solution to the problem enables the selection of aggregation points to well adapt to dynamic change of the network structure,and can dramatically reduce the communication cost and enhances the transmission reliability.Preliminary simulation results show that the OAPS algorithm efficiently reduces the data-collection cost by nearly 50% compared with other data collection methods without aggregation.