A major challenge for traditional systems is the lack of capabilities to automatically discover alternative solutions and actively deploy corresponding functions to intelligently adapt to changes in dynamic environments. In this paper, we continue our previous research of developing a two-layer architecture for modelling industrial cyber-physical systems and focus on the IEC 61499 function block based low-level physical module design. We propose an architecture model to integrate function blocks with intelligent agents to support self-management capabilities for low-level physical modules to quickly respond to changes. In the proposed architecture, a self-manageable service model is introduced for IEC 61499 function block modelled systems and designed as a multi-agent model with Self-Manageable Service Execution Agent, Self-Configuration Agent, Self-Healing Agent, Self-Optimization Agent, and Self-Protection Agent. The proposed modelling framework is tested with preliminary experiments on Raspberry Pi using the agent modelling tool SPADE and function block modelling tool Eclipse 4diac.