In the vision of the intelligent factory of industry 4.0, many enterprises are eager to increase the productivity by analysing the production data collected from equipment in runtime or products. Developing a continuous yield improvement system framework with big data capability shall gain an insight into the demands of continuous yield improvement. In this paper, we developed a cyber-physical-style continuous yield improvement system (CP-CYIS) with big data capability for providing a solution to rapidly construct the required yield improvement process for the shorter time on finding the root cause. The main idea is to divide a complete analytic flow into two part: the YI process describing only the flow without execution details and the YI action providing an analytic function for a YI process that involves the YI action. In this way, the development engineers can focus on the related modules that the process engineers concern. On the other hand, the executions of these actions are automated accomplished with assistances of the proposed system, so that the time spent on finding the root cause can be reduced. Finally, we deploy the CP-CYIS to a private cloud based on VMWare, and apply the CP-CYIS to a semiconductor factory for conducting integrated tests. Testing results of a case study show that the CP-CYIS has similar execution performance to the traditional solution. The development of this paper can provide a useful reference for industrial practitioners of the manufacturing industry to construct cyber-physical-style manufacturing systems.