At the tactical planning level in a multi-project environment, uncertainties are inherent to the workloads, and costs may be involved for using non-regular capacity and violating project due dates. We propose an approach to identify whether non-regular capacities might be needed to meet the projects' due dates. This problem is known as rough-cut capacity planning (RCCP) problem under uncertainty. We propose a possibilistic approach, which is based on modelling uncertain workloads with fuzzy sets. We present the resulting fuzzy rough-cut capacity planning (FRCCP), and show that we can use the possibilistic approach to provide a robust solution with a fuzzy resource loading profile that supports managers in decision making. We provide a simulated annealing approach to solve the FRCCP, and test it against several existing RCCP approaches. For the experiments we use real life instances from a shipyard maintenance centre.