Flood is an enormous amount of water overflowing in a river due to occurrence of unusual precipitation in the catchment or melting of ice in case of snow-fed rivers. River flow in delta stage is more prone to floods, where huge settlement of people and farming practices are common. Hence, prediction of flood extent is of great importance, especially for preparing emergency action plans and issuing early warnings. In the present study, a coupled 1D-2D (local-inertial) model is proposed to develop for river network and floodplain system. The governing equations are the 1D Saint-Venant equations for simulating the flow in the river-networks and the simplified local-inertial equation in the form of mass conservation for the flow over the floodplains. These river equations are discretized by using the implicit Preissmann scheme. The discretized equations are then linearized and solved by double sweep algorithm. The local-inertial 2D model will be developed utilizing the sub-grid concept so that micro-topographical features on the floodplains can be represented with higher accuracy. The 1D and 2D models will be coupled by transferring mass using a weir type equation. A number of coupling equations and strategy will be evaluated for better accuracy. The 1D river-network model will be validated with experimental observations and the coupled model will be validated with the observed flood water marks during a flood event. The 1D network model will then be applied to simulate flow in the river-network system of Kuttanad region in Kerala and Chennai where frequent flooding occurs.