The resource-constrained (single) project scheduling problem has been widely studied in literature. A common assumption in its basic version is that the execution of activities cannot be interrupted, that is activities are non-preemptable. Very little effort has been made to analyse the impact of activity preemption. In addition, classical project scheduling models only consider the case of a single project. This paper considers the problem of scheduling multiple projects under the assumption of activities preemption. A heuristic algorithm is proposed and its performance is analysed using several well-known priority rules. Computational experiments are carried out using the well-known datasets from the PSP Library. Results show the advantages of incorporating activity preemption when scheduling multiple projects.