This paper proposes a novel data prioritization and transmission mechanism to minimize the number of packets transmitted and reduce network overload using the constrained application protocol (CoAP) in resource-constrained networks. The proposed scheme adopts four classification parameters to classify and prioritize data from a sensor. With the packet prioritization scheme, the sensed data having the lowest priority is only delivered using the proposed keep-alive message notification to decrease the number of packets transmitted. The performance evaluation demonstrates that the proposed scheme shows the improvement of resource utilization in energy consumption, and bandwidth usage compared with the existing CoAP methods. Furthermore, the proposed scheme supports quality-of-service (QoS) per packet by differentiating transmission delays regarding priorities.