A roll surface in aluminum processing is smooth, has strong reflective characteristics, and is easily affected by ambient light when using an infrared temperature sensor to measure the temperature, resulting in low temperature measurement accuracy on the roll surface. The cooling control system affects the precision of cooling treatment on the roll surface, resulting in poor product quality. In this study, an infrared temperature compensation algorithm based on light intensity is proposed and constructed to improve the accuracy of ambient light measurement of the surface temperature of a strong reflector. Experimental results show that this method can compensate for measurement errors caused by changes in illumination intensity, thereby improving the measurement accuracy. The algorithm is simple and adaptable and provides a new approach to strengthening the accuracy of temperature measurement given speed change.