Software Defined Networking Technology is a network management methodology that allows for flexible, programmable efficient network configuration to boost network performance and monitoring. A network flow is defined as a sequence of packets from a source computer to a destination. The flows based on volume are broadly classified into two categories namely heavy-hitter or elephant Flow and Mice Flow. The Elephant Flows may be present in a small number in a network but they consume a subsequent amount of bandwidth of the network. This may lead to congestion, traffic unbalancing, etc. in network. So, classifying flows and scheduling them efficiently becomes an important service in a network nowadays. In this paper we have done an extensive survey of existing literature related to elephant flow detection in software defined networking and approached the detection of flows based on different machine learning model for prediction of volume rate and flow duration of network flow in which Random forest perform better in term of accuracy compared to others.