Abstract: Wireless technology has advanced tremendously over the past decade, introducing a wide array of devices with networking abilities. Adhoc networks are dynamically created and maintained by a group of wireless enabled devices (gargets) with out the assistance of pre-existing infrastructure like Base Station (BS) for communication purposes. The intrinsic attributes of adhoc networks, such as dynamic network topology, constraint battery power and quality, limited battery power, and large number of heterogeneous nodes, make very network management significantly more challenging than stationary and connected networks. In this paper, we proposed a dynamic topology management scheme, in which we consider the issue of performance enhancement of wireless networks in terms of �Power consumption of node as well as system, node density in the backbone routing and delay and node life time�. Minimizing power consumption is an important challenge in mobile networks. Wireless network interface is often a device�s single largest consumer of power. To fulfill the above requirements, we have designed a topology management scheme, which allows as many nodes of the adhoc network as possible to be in the power sleep mode, and most of the time they should be path between any source and destination to route the packets with minimal delay. In our topology management scheme, MARI( Mobile Agents with Routing intelligence) nodes are selected in such a way that, the MARI nodes have maximum power level among their one hop neighbors and all non-MARI nodes are with in the transmission range of MARI nodes. These MARI nodes have the routing intelligence i.e. they make all decisions relating to routing. The gateway nodes having sufficient power level are selected so that they can forward packets between MARI nodes. Gateway nodes do not have routing intelligence. These MARI and gateway nodes stay continuously awake to route the packets of other member nodes.