The main challenge for IoT-enabled networks lies in the transmission of large volumes of data produced by sensor nodes. The overconsumption of communication power presents a risk to the longevity of nodes. Hence, it is imperative to offer remedies for network-related problems such as Quality of Service (QoS), security, network heterogeneity, congestion avoidance, reliable routing, and energy conservation. Routing protocols are crucial for addressing the aforementioned issues of sending data across different firms. Data aggregation techniques are essential for the collection and consolidation of information in order to reduce traffic congestion , operational costs, energy consumption, and the lifespan of the network. Developing reliable, energy-efficient, and efficient route planning is challenging in scenarios where data is aggregated for Internet of Things (IoT) applications. The current investigation presents a novel routing system named Cluster-based Energy-aware & Nearest Neighbour Protocol (CEnNP), which utilizes NS2 simulation. The technique use decision trees and neural networks to evaluate the probability of a successful delivery. When training the model, we consider various factors such as the predictability value obtained from the CEnNP routing scheme, node popularity, node power consumption, speed, and location. The simulation findings indicate that CEnNP outperforms a reliable routing system for NS2 1 in terms of successful deliveries, lost messages, overhead, and hop count. However , these enhancements result in only marginal increments in buffer length and buffer occupancy levels. The proposed hybrid routing technique consists of two main stages: cluster formation and intra-and inter-cluster routing. The findings indicate that CEnNP outperforms prior studies in terms of network resilience, packet transmission efficiency, end-to-end delay, and energy consumption