In order to meet digital, high-accuracy detection requirements, a measuring system of infrared sight zero variable quantity based on charge coupled device (CCD) machine vision is designed. A new zero variable quantity measuring model is established. According to measuring system technical indicators and requirements, the off-axis parabolic collimator and pinhole target are designed. To reduce infrared sight zero variable quantity measuring error caused by camera tilting, a camera pose adaptive correction model is established, which is firstly applied to infrared sight zero variable quantity measuring. An image interpretation software is designed based on visual studio software development platform, using Gravity method to locate spot center of target image, and using Zernike orthogonal invariant moments to classify pixel on the edge. A functional prototype is successfully developed, processing equipment is tested after adjustment. The results show that its detection uncertainty is better than 0.02 mil, which can avoid measuring error caused by camera tilting and satisfy digital, high-accuracy test requirements on infrared sight zero variable quantity.