Abstract: As a key cooling equipment for large-scale industrial facilities in power plants, the operation status of the air cooling island directly affects the efficiency and safety of the unit. The traditional manual inspection method has problems such as low efficiency, discontinuous data, and poor real-time performance. This article proposes an intelligent temperature measurement system for an air-cooled island based on a track based inspection robot and infrared thermal imaging technology. By equipping the robot with a high-precision infrared thermal imager and combining AI data analysis, real-time monitoring and freezing warning of the temperature field on the air-cooled island are achieved. Practical application has shown that the temperature measurement accuracy of the system reaches ±2 ℃, and it can identify abnormal freezing on the surface of the heat transfer tube bundle, covering more than 95% of the visible range, significantly reducing the cost of manual inspection, and providing important basis for the start stop control and freezing treatment of the air cooling island fan.