[en] A new method of detecting thermal anomalies using in infrared camera has been developed for remote plant machine monitoring. Anomalous phenomena, such as an air flow shortage of the cooling system in an electric motor, were successfully detected at an early stage by this method even it the motor operation state is thermally transitional. This method is based on the simple principle of comparing thermal image data between pre-stored reference thermal image obtained during the normal machine operation and the monitored images. The images are taken by the infrared camera. A thermal anomaly index is calculated as a two-dimensional cross correlation coefficient from the two images. Experiments were carried out to confirm the usefulness of the proposed method. In the experiments, a pseudo-thermal anomalous occurrence was generated by closing the air inlet of the motor cooling system. Reference thermal images were taken at the rated load condition and the pseudo thermal anomaly was successfully detected by monitoring the anomaly index. It was also confirmed that thermal conditional change associated with normal motor operation, such as load change, gave very little variation to the index value and a normal judgment was made, although the load change resulted in a significant temperature change on the motor surface. This method is suitable for not only the detection of anomalous phenomena on rotating machines, but also the detection of thermal anomalous phenomena on several kinds of plant machines