During the execution of manufacturing processes, problems arise and they have to be solved systematically to reach and exceed production targets. Normally, a production team analyzes and solves these problems, with the support of different methodologies and working directly on the shop floor. This paper presents an ontology-based approach to easily capture and reuse the knowledge generated in such a process of Manufacturing Problem Solving (MPS). The proposed ontology is used as basis in an ad-hoc MPS software system. The architecture of the MPS system is based on the integration of three technologies: PLM (Product Lifecycle Management), CBR (Case-Based Reasoning) and software agents. The PLM system is used as an automatic source of the problem context information. The CBR system is used as repository of cases and artificial intelligence tool to support the efficient reuse of knowledge during the resolution of new problems. A software agent platform allows developing an integrated prototype of an ad-hoc software system. This paper shows the architecture of the MPS system prototype.