With the increasing demand for individualization, new smart manufacturing models are implemented in order to adapt to the market demands of high quality, multi-variety and fast delivery. However, new problems in production and operation are urgently needed for research. It is especially significant to study the automatic flow workshop planning based on data-driven modelling and simulation. Firstly, according to the rapid analysis of various data of the workshop, UML-based object-oriented data model is formulated. Then the application module of data-driven simulation modelling is established by using the simulation platform. The bottleneck recognition based simulation optimization module for the tabu search algorithm with guiding rules is designed. Data-driven modelling makes the model more flexible and reusable which can run a variety of different experiments by modifying the external data to drive simulation model. Finally, experiments are conducted to verify the effectiveness of the simulation optimization method proposed in current study. In addition an optimized production line configuration plan is provided.