Data aggregation is a process of collecting data in an energy-efficient manner by which the life span of the network is enhanced. Aggregating data is a process of comprising the transmitted packet, in the sense the packet's only necessary information is sent to the remote sink node for further processing. While designing efficient data aggregation algorithms, a few extra requirements have to consider such as energy capabilities of sensor devices, energy resources, and computational capabilities. A novel Statistical Indexed X-means Clustering-based Divergence Butterfly Optimized Recursive Deep Learning Network (SIXC-DBORDLN) is introduced for aggregating the sensed data at the sink node with higher accuracy and minimal time. In SIXC-DBORDLN, Recursive Deep Learning Network comprises five layers, namely one input layer, three hidden layers, and one output layer for performing the data aggregation in WSN. In the SIXC-DBORDLN technique, a number of sensor nodes are considered as input in the input layer. Robin Hood indexive Hannan–Quinn informated X-means clustering algorithm is applied to generate ‘x’ number of clusters based on energy level for the first hidden layer. The residual energy value is calculated for every sensor node in the cluster. Then, assign the sensor nodes to the cluster whose residual energy value is closer to the centroid value. The cluster centroid value gets updated and the process gets repeated until all sensor nodes get reassigned. In hidden layer 2, the cluster head is selected using Jensen- Shannon divergencive butterfly optimization for every cluster based on the residual energy. The sensor node with higher fitness is taken as the cluster head. In hidden layer 3, the cluster head collects the data from all sensor nodes and sent to the sink with minimal delay. In the output layer, the cluster head transmits the data packets to the sink node. In this way, the energy-efficient data aggregation is carried out in WSN. Experimental evaluation is carried out using an agriculture dataset on the factors such as energy consumption, delay, packet delivery ratio, and data aggregation accuracy with respect to a number of sensor nodes.