Mining negative itemset and association rules is always an interesting but tough because of its high search space and complexity. Almost all the methods proposed for negative association rule mining finds negative itemset along with positive itemset. It is difficult to find meaningful association between items by defining various kinds of correlation constants along with the formal support confidence frame work as all the itemset are not interesting. So far there are several algorithms proposed successfully to find positive regular itemset but no algorithm exists to mine negative regular itemset. Hence in this paper we are presenting an efficient NPRISM algorithm to mine both positive and negative regular itemset. We opted vertical format in this algorithm in order to improve the efficiency and for convenience. The experimental results show that our algorithm successfully finds all the positive and negative itemset by scanning database only once.