Data is increasing rapidly day by day along with the transactional database. Dividing this data and storing it in a distributed manner is an effective way for storage and retrieval. Mining such distributed data with minimum dependence between sub-problems is a crucial task. Finding frequent itemsets and corresponding association rules is a big challenge while considering the aggregation in a distributed environment. To overcome these challenges, we propose a distributed frequent itemset generation and association rule mining algorithm using MapReduce programming model. The proposed scheme generates frequent itemset and mine association rules using a synthesized distributed technique. The rules are mined in a distributed manner, and then weights are assigned to subsets of data and association rules. A proper mixture of association rules that are generated in distributed manner is done using a weighted approach. This paper presents a novel MapReduce-based synthesisapproach, which can work well over a distributed storage of large amount of data.