Abstract: Aiming at the problem that the current two-point non-uniformity correction method was difficult to track the non-uniformity of infrared image drift with time in infrared search and track system(IRST) real-timely, a novel real-time joint correction algorithm based on the improved two-points correction method and scenes was proposed. This algorithm uses the two-point method to get the basic correcting parameter, then analyzes mass real-time image data to find the values related to the system non-uniformity changing with time. The degradation of infrared image caused by two-point non-uniformity correction is solved and intensive tests validate that the relative non-uniformity decreases from 5% to 2% after using the joint correction method, it has good time stability and better correction effect.