Condition monitoring is essential for the OEE of machine tools. Existing solutions are customized to specific settings. However, linear guidance systems commonly used in machine tools are exposed to varying process conditions. Thus, this contribution proposes a concept for a transferable condition monitoring system, which enables a static system to be applied to different settings. The solution is composed of a combination of data preparation methods, feature generation and an anomaly detection model. The system is demonstrated on two test beds with different linear guidance systems. The selected isolation forest for anomaly detection is trained on a series of experiments from one test bed before transferring the condition monitoring to the other test bed.