This paper concerns the application of the maintenance task scheduling approach in a distributed context. We are particularly interested in the case where there is, on the one hand, a central maintenance workshop (CMW) in which therepair cycles are carried out and, on the other hand, a mobile maintenance workshop (MMW) for repair by replacement at several sites according to a predefined schedule. The maintenance costs are then partly linked to the costs of spare parts, travel, repair or replacement times. A maintenance cost optimization methodology based on the use of meta-heuristics is proposed in this article. A comparative study was carried out using two optimization algorithms: the genetic algorithm and the simulated annealing. The algorithms were applied to a company in the Rhône-Alpes region.