In this paper, a smart hybrid model for shop floor simulation and control is proposed based on dynamic modelling and control theory’s analytical tools. The model is smart because the production levels are automatically adjusted based on feedback loops, considering shop floor current states (i.e., it is a closed-loop model). It is hybrid because it deals with discrete and continuous transitions, and integrates two approaches: switched arrival systems (SAS) and bond graphs (BG). This novel integrated SAS + BG model is applied to a production system in the food industry and four production control rules (well-known from the operations management literature) are tested: pure push and pull rules and extended push and pull rules. The results showed that the SAS is effective to dynamically represent intermittent production systems, and its use, coupled with bond graphs, allows the definition of capacity adjustment and production control rules that are effective to control the work in process (WIP). The extended pull rule outperformed the remaining ones, corroborating prior literature. This paper contributes with a novel approach that is modular and can be extended to represent the closed-loop dynamics of different configurations of multi-product multi-station production systems, with a disaggregated level of detail.