The work aims to study online luggage stacking algorithm to solve the problem that most of the 3D boxing algorithms cannot solve the problem of online airport luggage stacking under the background of air transportation. First of all, a new sequential constraint and online multi-size luggage stacking mathematical model was proposed based on the practical application environment. Then, a rule that was different from traditional online question which only obtained one item at a time was designed. The position of 3D camera for capturing luggage information was changed to get more luggage information in advance. An online heuristic ''packing point'' algorithm was proposed to obtain K luggage size information and construct a combined stacking strategy each time. Finally, 3 pieces of real airport luggage information were gathered at a time for simulation and experiment. The results proved that the stacking layout scheme given by the algorithm could not only ensure the efficient stacking of checked luggage of different sizes in online state, but also obtain the packing rate of 84.86% by the optimized algorithm. Meanwhile, it had other good properties, such like high stacking stability. The stacking layout scheme given by the algorithm is reasonable and has good filling rate. It provides a new direction for the algorithm design of online 3D boxing under the background of air transportation.