Abstract: According to the technical traits of Network Centric Airborne Defense Element, an aim point selecting algorithm based on sparse pixel vectorization was presented. On the basis of summarizing the structural and infrared characters of ballistic missile during the boost and ascending phase, the target's plume was segmented by extending the global histogram threshold method to a dual threshold method with the histograms for different areas, then the sparse pixel vectorization was used to find the center axis of the target's infrared image, and the aim point was found in two ways according to the proportion between the skin and the plume, finally the aim point was confirmed with the consistency of direction and distance in the subsequent frames. Experiments show that the algorithm presented here is able to realize the aim point selecting for infrared imaging endgame guidance fast and adaptively without manual work or traversing all the pixels.