Mobile adhoc networks are a gathering of mobile nodes that communicate through a wireless channel without any unchanging infrastructure such as an access point or a mobile base station. In this network, every node transmits data from the forwarding packets to its neighboring nodes. Some nodes may not participate in forwarding the packets for the reduction in its resources. The presence of selfish behavior among the nodes may lead to network partition and makes a major harmful impact in the network operation. To avoid such conditions, selfish node inference is essential. To overcome this problem some techniques like, a credit based technique, reputation based technique, Record and Trust Based Detection techniques (RTBD) and so on, have been used. The proposed model provides security to the neighbor nodes by using pseudo random number generation using RSA algorithm, to avoid the neighboring node from being compromised by the selfish node. It significantly improves the performance metrics such as, packet delivery ratio, energy, and routing overhead.