Time-Dependent Origin-Destination (OD) matrices are a key input to Dynamic Traffic Models, microscopic and mesoscopic traffic simulators are relevant examples of such models, traditionally used to assist in the design and evaluation of Traffic Management and Information Systems (ATMS/ATIS). Dynamic traffic models are also starting to be used to support real-time traffic management decisions. The typical approaches to the time-dependent OD estimation have been based either on ad hoc heuristics using mathematical programming approaches, or on Kalman-Filtering. The advent of the new Information and Communication Technologies (ICT), as for example Automatic Vehicle Location, License Plate Recognition, detection of mobile devices, Vehicle to Infrastructure (V2I) and so on, makes available new types of traffic data of higher quality and accuracy allowing for new modeling hypothesis leading to more computationally efficient algorithms. This paper extends the previous research on Kalman Filtering approaches for Freeway OD estimation using these data, to more complex topologies of urban networks were alternative path choices between origins and destinations are available. Ad hoc procedures based on Kalman Filtering have been designed and implemented successfully and the numerical results of the computational experiments are presented and discussed.