The present work deals with a dynamic grouping maintenance strategy for complex systems whose structure may lead to both positive and negative economic dependencies. Impacts of the system structure and preventive maintenance durations on maintenance grouping are investigated. To find the grouped-maintenance plan, a dynamic maintenance approach based on a rolling horizon scheme and a genetic algorithm is proposed. Moreover, a dynamic situation in which time-limited opportunities may occur is studied and integrated in the proposed dynamic maintenance approach. Finally, a 12-component system is used to illustrate the performance of the proposed maintenance grouping strategy.