Daily life applications of machine-to-machine (M2M) communication are constantly increasing. Typically, M2M communication systems comprise numerous small, cheap and autonomous devices that communicate with each other while monitoring environmental conditions. After their deployment, however, remote entity management of a node is still needed for firmware or software updates required for bug fixes, functional changes or other maintenance. Therefore implementing remote entity management for M2M service capability in third generation partnership project machine-type-communication is a major challenge. Previous works have attempted to reduce traffic by updating only the software changes in the M2M device. However, the device must still reboot after a software update. Rebooting the device is costly since the previous runtime states are lost. The devices expend time and bandwidth when synchronising with other nodes and when rebuilding the routing table. Hence, the dynamic software update model (DSUM) proposed in this study is designed to enable remote entity management for M2M. A dynamic software update programming model and a prototype implementation for M2M service capability are also presented. By allowing M2M service capabilities to update the software without rebooting the node, the DSUM preserves precious runtime states. The tests in this study showed that software updating required only 88 clock cycles. This framework not only enables dynamic replacement of remote entity management functionalities at runtime, it also reduces power consumption by avoiding the need to rebuild the network topology for devices in an M2M communication network.