Dealing with a large amount of data makes the requirement for efficient data accesses more critical in Data Grid. Through improving performance of data replication, job execution cost is reduced. In this paper, we proposed a new strategy of replication based on organizing the data in Data Grid based on its property. Particularly, we organized the data into several data categories that it belongs to and this categorization helps improving data replication placement strategy. In addition, the paper also introduces Dataset Scheduler which helps optimize obtaining input data for job. The performance studies are conducted using simulation tool and prove the improvement of scheduling performance over current approaches.