To solve the wide area measurement system(WAMS)mass data processing problems such as data redundancy and low processing efficiency,a cloud computing platform based on Hadoop is designed and implemented. The structure of this platform is described first. Then,a WAMS mass data loading method is designed based on Hadoop distributed file system (HDFS)and parallel data extraction-transformation-loading(ETL)for multiple file processing by using MapReduce. MPApriori data mining algorithm combined with MapReduce is proposed to discover the interplay of power sites when cascading failures occurred. Finally,through the regional network WAMS actual data processing,the effectiveness of mass data processing on Hadoop is proven. This platform is suitable for mass power grid files data mining by high performance local area network connection of a computer cluster.