A mobile ad hoc network (MANET) is a self-organized system comprised by multiple mobile wireless nodes. The node misbehavior due to selfish reasons can significantly diminish the performance of MANET. A selfish node attempts to use the resources only for its own purpose and it hesitates to share the resources with their neighbors. So, it is very important to detect the selfish nodes to improve the performance of MANET. Initially, an architectural model of a MANET is constructed and the communication between the mobile is originated. The packet drop can happen in MANET due to the selfish node or network congestion. In this paper, a Record-and Trust-Based Detection (RTBD) technique is proposed to detect the selfish nodes efficiently in MANET. The main reason for using trust in this analysis is to accelerate the detection of misbehaving nodes. This study has been carried out in order to analyze the detection of selfish nodes on essential network functions such as routing and packet dropping. The results show that the proposed selfish node detection method is very efficient, since the detection time of selfish nodes is diminished and the overall overhead is very low. The simulation study demonstrates that the proposed RTBD method enhances the selfish node detection ratio, packet delivery ratio (PDR), and average packet drop ratio.