The aim of this paper is to explore a new approach to obtain better traffic demand (Origin-Destination, OD matrices) in case of dense urban networks. To improve the global process of OD demand estimation, this paper is focussing on a new methodology to execute the dynamic OD matrices determination adapted to urban area. An iterative bi-level approach will be use to perform the OD estimation. The Lower level problem will determine dynamically the utilisation of the network by vehicles using heuristic data from traffic simulator. This simulation will be mesoscopic and a particular calibration will be done, focusing mainly on flow and route choice indicators. “Design of Experiment” will be use to evaluate the influence of the parameters and to assess the best parameters combination. The Upper level problem will proceed to an OD adjustment using Kalman filtering technique.