We describe in this paper the characteristics of remote maintenance systems and set up their bases. We also present its basic principles through a general structure model of the system to be managed. A heuristic-based optimization algorithm is proposed to solve the online preventive maintenance resources allocation problem for a distributed system based on the remote maintenance systems' generic architecture and given an availability target for a distributed system. This algorithm allocates the resources to minimize the cost of discharge of preventive maintenance tasks by considering precedence constraints among consecutive standby tasks and their emergency. The criteria for job emergency and priority were used to solve this problem while scheduling the right order tasks. Tasks are allocated according to resource availability along with load balancing. Computational results illustrate the efficiency of the approach implemented on different system configurations.