A particular challenging task in the context of the 4th Industrial Revolution consists in increasing production system automation in European small and medium enterprises (SMEs). However, automation – often addressed with robotization – represents major investments that are not always accessible for SMEs, in which small batch production requires change in an efficient manner. In fact, software programming and hardware (in terms of reachability and maximum payload) are the flexibility barrier of current industrial robots. On the other hand, researchers and industry are currently pushing their efforts to develop modular robots (in terms of hardware and software), which are task-specific computable and changeable. In this context, this paper aims to develop a new design methodology combining both knowledge representation for task definition and computational tools for structural and logical design using modular units. To demonstrate its suitability in an industrial context requiring reconfiguration capabilities, specific developments have been made and tested with a simple case.