Global competition has led to more complex production systems, in which efficient maintenance is critical for operational competitiveness. The effects of poor maintenance levels can have huge economic impacts in such production systems. To avoid this problem, it is necessary, on one hand, to be able to estimate maintenance needs in advance, thus be able to avoid unforeseen breakdowns and production interruption. On the other hand, management and planning of spare parts supply chain systems become more important due to the difficulty of ensuring spare parts availability while keeping reasonable inventory and transportation costs. To address both points, research in the domains of Intelligent Maintenance Systems (IMS) and Advanced Planning Systems (APS) for spare parts supply chains have been arising in recent years, providing means to forecast device failures by the analysis of sensorial inputs, resulting in the ability to forecast maintenance and spare parts needs more precisely. The integration of IMS and APS makes it also possible to adapt the machine parameters according to the reactivity of the spare parts supply chain and to enable its usage until the spare parts are available. One important challenge in this context is the proper integration of both types of systems, coping with the different backgrounds and levels of abstraction found, like shop-floor IMS devices and the systems employed by APS providers. This paper will address this integration problem by proposing an architecture to integrate IMS devices and APS planning components, starting from modelling domains and actors, analysing the effects variables and components from IMS and APS have on each other and conceptualizing an architecture for the integration. Expected results are high-level concepts for the system architecture, an ontology for modelling the spare parts and maintenance systems domain, proposition of an integration layer and features for message exchange and adaptive machine behaviour.