Abstract: This paper presents a method for predicting the minimum resolvable temperature difference (MRTD) of thermal imager based on the perceived threshold signal-to-noise ratio of the human visual system. According to the frequency spectrum theory and two-dimensional modulation transfer function of infrared thermal imaging system, combined with the concept of matched filter, the method describes the spatial frequency spectral function of 4 standard bars. After the 4 standard bars are modulated and transferred by the infrared thermal imaging system, the perceived signal-to-noise ratio of the human visual system can be researched by using the spatial frequency spectral function. A new MRTD model based on the perceived threshold signal-to-noise ratio of the human visual system was established. The visual range prediction of thermal imager was achieved based on the MRTD model and visual range model. Simulation results show that the predicted visual range agrees with the experimental data.