By general admission, image fusion technology can improve target detection. However, research on quantitative analysis for predicting target’s detection probability of image fusion system is lacking. To solve this problem, we propose a computational model for calculating target’s detection probability of visible and IR image fusion system. This model is designed based on five components: target’s own spectral contrast to its background; visible and IR detector’s characteristic which is described as the spectral matching degree between detector’s and target’s spectral distribution; environment illuminance condition; fusion image local quality defined by local target contrast and local sharpness; distance component based on target’s distance and size. The proposed model is employed in two different detection tasks. Experimental results show that this model is meaningful and effective on predicting target’s detection probability of visible and infrared image fusion system since it corresponds well to the perception evaluation.