In this study, a spatial-noise subdivision evaluation model of an infrared detector is introduced. The spatial noise was classified into five independent noise components according to their graphic features. A frequency domain analysis method was used to accurately separate the spatial noise components and calculate their spatial noise equivalent temperature difference. By testing the spatial noise components of the mainstream uncooled infrared detector, we found that the main components were high and low spatial frequency noise. The effects of spatial noise on the performance of the infrared system were also analyzed for two typical applications of imaging and temperature measurements. We found that spatial high-frequency noise is the main factor affecting the minimum resolvable temperature difference performance of the infrared imaging system, whereas spatial low-frequency noise is the main factor affecting the accuracy of temperature measurements in the infrared temperature measurement system. The results indicate directions for further studies on the spatial noise of uncooled infrared detectors.