電子零組件廠商採購主要發生於電子廠商與供應商間之買賣行為,然而多數電子廠商對於供應商的評選、協商,依舊是採用人工作業,即使有更多供應商可以選擇,卻因為缺乏時間與人力,導致其效益與效率降低。本研究藉由文獻回顧與專家訪談,彙整出電子產業的採購流程、協商策略之現況及可協商的議題,並發展出一個以智慧代理人為基礎的協商架構,以解決上述問題。本研究由第三仲介者的角度來建構交易平台,並提出一個符合多個買方與賣方且多議題的交易環境,買賣雙方透過此平台同時指派多個代理人程式互相協商。在協商過程中,為了能符合人類的思考模式及反映實際世界的協商情況,本研究運用模糊理論與互惠合作動態策略以調整讓步,以提高協商效率、買賣雙方交易的可能性及滿意度。在本研究所假設的交易環境下,建置一系統雛型,以驗證協商機制之效益。 The purchase behavior takes place mainly between electronic manufacturer and its supplier. However, the activities of negotiation and decision in electronic manufacturer with its supplier are still using manual operations, due to the restriction of limited time and manpower, even if there are many options of suppliers, the benefit and efficiency is greatly reduced. In this research we compile the purchase process of electronic industry, the current conditions of negotiable strategy and the negotiable subject by reference and an interview. We also develop a B2B e-Marketplace with the basal intelligence to solve the above problem. This research constructed a trading platform of the third party’s view, and put forth a trading environment to conform several subjects, buyers and sellers. The business of both parties makes use of several agent programs to negotiate with the platform at the same time. For the sake of thinking mode according to humans and reflect the negotiation condition of actual world in the process of negotiate. This research use fuzzy theory and cooperation strategy to adjust concession by dynamic state to raise negotiation efficiency, to increase the possibility of success and satisfaction between the businesses of both parties.Under the trading environment proposed by this research, we build one system model to test and input the negotiation information into the system, and verify the benefit of the negotiation mechanism.