In this paper we present RDPPlan, an automated problem solver which generates plans of actions in order to satisfy logical goals and numerical goals on uncertain resources. Planning with resources is a component of many applications which range from robot planning to automated manufacturing and automatic software composition. In the classical planning model the actions describe purely logical state transitions; some extensions have been recently proposed in order to manage numerical resources which can be produced/consumed in exact amounts. Unfortunately in real domains it is impossible to make accurate and exact previsions about resource production/consumption because of the inherent uncertainty of real world. The planning model introduced in RDPPlan allows to manage uncertainty about the initial value of resources and actions that can make uncertain updates of numerical resources. The proposed model uses the notion of trapezoidal fuzzy intervals to handle the uncertainty on resource values; the solving algorithm extends, for fuzzy resources, the propagation rules of the planner DPPlan.