Cloud computing is a new emerging model in the field of computer science. For varying workload Cloud computing presents a large scale on demand infrastructure. The primary usage of clouds in practice is to process massive amounts of data. Processing large datasets has become crucial in research and business environments. The big challenges associated with processing large datasets is the vast infrastructure required. Cloud computing provides vast infrastructure to store and process Big data. Vms can be provisioned on demand in cloud to process the data by forming cluster of Vms . Map Reduce paradigm can be used to process data wherein the mapper assign part of task to particular Vms in cluster and reducer combines individual output from each Vms to produce final result. we have proposed an algorithm to reduce the overall data distribution and processing time. We tested our solution in Cloud Analyst Simulation environment wherein, we found that our proposed algorithm significantly reduces the overall data processing time in cloud.