A Mobile Ad hoc Network is a wireless network in which a central management scheme is not available; this causes nodes to dynamically self-organize each one playing both router and node’s roles. Compared to the other types of wireless networks, the infractureless nature of this type of network makes it provide various advantages with some numerous problems of course. One problematical approach toward routing in MANET is the security issues, which, combined with reliable data transfer problems make some MANET’s users unsatisfied. This security problem is explained in the fact that as participating nodes regularly communicate and exchange information; there must be a level of trust between them. Some weak protocols are designed in such a way nodes randomly exchange messages without any control due to the lack of trust management mechanisms which may be required to prevent such unsecure information exchanges; thus opening a way to unwanted malicious nodes and intruders compromising the security in MANET, and in turn, resulting in unreliable data transmission. This paper conducts a performance evaluation of various security-aware algorithms based on trustworthy schemes namely trust evaluation with direct observation and trust evaluation with indirect observation. These two schemes are evaluated against the existing one while varying the routing metrics available for MANETs; throughput, hacked packet ratio, end-to-end delay ratio, and the overall trust valuation metrics. This performance evaluation is conducted for the Optimized Link State Routing Protocol version2 (OLSR v2), the successor and enhancement of OLSR v1. It has been chosen thanks to its capability to efficiently extend MANET’s neighborhood discovering process, a feature which makes it easy to detect malicious neighbor and distant nodes in the network. The simulations are carried out using NS-2 while varying the number of nodes, the outcomes show that the trust evaluation with indirect observation approach outperforms the other schemes; the direct observation scheme mediumly performs while the existing one is always worsened except with the end-to-end delay parameter.