In recent years, we are constantly facing the increase of Internet of Things devices in the fields of health, agriculture, industry and other APPLICATIONs. The using of these networks will increase by improving the quality of service. Usually, data are generated heterogeneously by IoT’s sensors at different times with different sizes and priorities, and none of previous works in this field do not consider all these modes together; on the other hand, these studies either have worked on the queue model and priority package selection, or on the channel ACCESS methods and its prioritization. Therefore, in our proposed model in this paper, with the aim of improving quality of service, thight scheduling of delay-sensitive packets, and also avoiding starvation for lower priority packets, we consider an architecture including two levels of prioritization. In the first level, the packet is selected for sending based on a type of dynamic prioritization and relative to its delay deadline field. Since in this paper the APPLICATION of Internet of Things in health care is considered, the sensors generate data both of modes, periodically (fixed) and on demand (critical), and therefore, two queue models D /G/1 and M/G/1 are used. In addition, the second level includes channel prioritization. At this level, by using a control channel and setting the waiting time of the sending node for listening to the channel, we have sending the packet through the control channel. Furthermore, a consolidated channel hopping model has been used for nodes to use the available bandwidth. Also, a buffer to keep the information of the last connection of the node helps to send and receive faster with less number of switching. After simulating and comparing this model with the 802.11 standard and similar media ACCESS protocols, significant improvements in increasing the packet delivery rate, operational throughput, and also reducing the end-to-end delay are showed.