In this paper we describe a real-time rolling stock rescheduling model for disruption management of passenger railways. Large-scale disruptions, e.g., due to malfunctioning infrastructure or rolling stock, usually result in the cancellation of train services. As a consequence, the passenger flows change, because passengers will look for alternative routes to get to their destinations. Our model takes these dynamic passenger flows into account. This is in contrast with most traditional rolling stock rescheduling models that consider the passenger flows either as static or as given input. Furthermore, we describe an iterative heuristic for solving the rolling stock rescheduling model with dynamic passenger flows. The model and the heuristic were tested on realistic problem instances of Netherlands Railways (NS), the major operator of passenger trains in the Netherlands. The computational results show that the average delay of the passengers can be reduced significantly by taking into account the dynamic behavior of the passenger flows on the detour routes, and that the computation times of the iterative heuristic are appropriate for an application in real-time disruption management.