In recent years, energy consumption and data gathering is a foremost concern in many applications of wireless sensor networks (WSN). The major issue in WSN is effective utilization of the resource as energy and bandwidth with a large gathering of data from the monitoring and control applications. This paper proposes novel bandwidth efficient cluster based packet aggregation algorithm (BECPA) for heterogeneous WSN. It combines the idea of variable packet generation rate of each node with random data. The nodes are randomly distributed with different energy level with equal in numbers. It uses the perfectly compressible aggregation function at cluster head based on the correlation of packets and data generated by nodes. The aggregation functions prevent transmission of redundant packets. Compare to state-of-the-art solutions, the algorithm shows significant energy saving by reducing the number of packets at the sink with better bandwidth utilization in packet aggregation than data aggregation.