In this paper, a three-dimensional noise model is applied to analyze the noise characteristics of a gaze-type mid-wave infrared detection system, and the suppression effects of information processing methods on various types of noise components are investigated. The image data are continuously collected and analyzed at different integration times for different temperatures of surface source blackbodies, and it is found that when the integration time and the incident energy of the blackbody increase, each noise component decreases rapidly and then tends to be constant in a nonlinear form, and each noise component presents its own unique image pattern. It is also found that spatial noise is the main component of system noise in the original image without any information processing. After comprehensive application of spatial and temporal noise suppression methods, the spatial noise is almost completely suppressed, the temporal noise is reduced by about half, and the transient pixel noise becomes the main component of noise.