The moisture content in lubricating oil is one of the most important factors to reflect the health and effectiveness of it, thus to monitor the moisture content in lubricating oil in real time is crucial for high speed rail. In this paper, we developed a compact moisture-content monitoring system based on the visible-near-infrared (Vis-NIR) spectroscopy technology, which was shown to be able to determine the moisture content in lubricating oil in a fast, simple and accurate way. In this system, a reflection mode optical probe was developed for sending and receiving optical signals through the viewport of gear box. By employing the reflection optical probe, one can measure the spectrum of the lubricating oil in a fast way by simply putting the probe on viewport of gear box, and therefore has the potential to realize the real-time monitoring of the status of the lubricating oil. In order to verify the feasibility of the system, both reflection and transmission spectral of lubricating oil were taken. Partial least square regression (PLS) and back propagation neural network (BPNN) algorithms were used to establish the processing model. Modelling results show a good agreement in between two different probing modes. As a result, the effectiveness and reliability of the system have been proved, which provides a simple yet accurate method for real time monitoring the healthy status of the lubricating oil for the safe operation of the high-speed rail. © 2020 Elsevier GmbH