<p>A perfect auction policy is one of the most strategic elements that contribute to success factor for any e-Procurement system. An auction policy can be only term as an effective if it really offer win-win situation to both the bidder as well as to the merchant. After reviewing existing studies on e-Procurement system, it is found that there isno effective research work focusing on this point and maximum research contribution has limited its scope to certain application or case studis. Hence, the proposed system introduces a novel e-Procurement system which is equipped by an itelligence-building process for performing predictive analysis of ongoing auction process. A mathematical modelling is implemented where all teh variables have been formed using practical implementation of auction system and followed by optimization process using regression-based approach. The study outcome shows that proposed system offers better response time and higher predictive accuracy in contrast to existing approaches.</p>