Mobile Ad Hoc Network (MANET) is one of the rising wireless configurations which have self-organizing wireless mobile nodes and adaptive network connectivity. Because of its dynamic nature and constantly changing topology, security is a critical issue for MANET. So, a trust-based secure energy-efficient routing protocol has been suggested in MANET which predicts the node’s trust level based on the static threshold value to detect malicious nodes and choose the best paths for data transfer. But, this threshold might not be ideal regarding identification accuracy and Packet Loss Ratio (PLR). To solve this issue, a highly suitable trust level must be estimated according to the node parameters like mobility, Node Degree (ND), etc. Hence, this article proposes an adaptive Trust Threshold (TT)-aware secure energy-efficient protocol by considering the network parameters to adaptively choose the TT. The TT in this protocol is adjusted dynamically at each node based on the different network factors. Initially, these various network variables are determined for all nodes. After that, an Artificial Neural Network (ANN) classifier learns such parameters to identify the TT, which aids in predicting the node's trust level. Further, malicious and benign nodes are discovered properly based on their trust level. At last, the simulation results exhibit that the proposed protocol achieves 92.1% accuracy while deploying 300 nodes compared to the other protocols.