Centrality measure is a tool to identify key nodes/edges in networks having higher importance/control/load etc. Betweenness centrality ranks nodes to locate the nodes with maximum control of information flow or nodes heavily loaded due to information flow happening through the shortest paths. The amount of traffic passing through a node is estimated by computing its betweenness centrality. The highest value of betweenness centrality observed among all the nodes in a network is termed as maximum betweenness of the network. We consider the problem of finding a missing link in a network, adding which minimizes the network’s maximum betweenness. We propose a Graph Neural Network (GNN) based approach to predict such a missing link. Experiments are conducted on various types of synthetic networks to evaluate the performance of the proposed model.