During urban freight distributions, the use of an initial routing plan, although necessary, is by no means sufficient to address unforeseen events that are likely to occur during delivery execution such as vehicle breakdown, traffic congestion, and adverse weather conditions. Recent advances in telematics and positioning systems enabled freight carriers to dynamically monitor their fleet through vector maps and to improve relevant delivery network performance by using a posteriori reports that include various details concerning the execution of the distribution plan by every vehicle. Although the use of such technologies supports improved utilization and management of the delivery fleet, the related systems and algorithms are not typically designed to address unforeseen events in a systemic fashion. As a result, interventions are often performed manually (for example, through voice communication between drivers and the logistics manager) and the resulting decisions are local with sometimes limited effectiveness. This paper proposes a real-time fleet management system for dynamic incident handling in urban freight deliveries. The system monitors distribution vehicles in real-time and recommends intelligent re-routing interventions in case of vehicle delay or vehicle breakdown so that two objectives are met: minimization of delivery delays and maximization of number of served clients. The system was tested within real-life operations of two freight carriers in Greece and the results demonstrated reductions in time window violations and increases in the total number of served clients.