Sandwich-like structures can significantly reduce the self-weight and size of load-bearing elements. However determining the flexural response of multi-layered sandwich structures is not a trivial task, particularly their buckling response. This study consists of novel theoretical work for extending Allen’s and Southwell’s buckling theories in order to develop a comprehensive model for the buckling response of multi-layered composite sandwich struts. The new model is applied here to predict the buckling of composite glazing, made of glass face-sheets separated by and adhesively-bonded to inner core profiles, thereby reducing the weight and thickness of conventional glazing panels. Existing physical test data on composite glazing is currently limited to flexural responses under out-of-plane loading and no published data exists for in-plane loading. Therefore three composite glazing struts were fabricated and tested in axial compression in this study, and the test data is used to validate the new model. As expected, failure of the composite glazing struts originated at regions of high shear stress close to end supports. It was found that the critical buckling loads predicted by the new analytical model are within 4% of those obtained from Southwell plots of the experimental data. Similarly, the new model provides a very good fit to the load-deflection and the load-stress responses from the experiments. This is a significant improvement on traditional Euler-based layered or monolithic models that, for this particular application, produce an underestimation of 90% and an overestimation of 40% of the critical buckling loads, respectively, and would therefore result in uneconomic or unsafe designs, respectively.