With the development and initial use of reusable spacecraft, online monitoring of the structural response of reusable spacecraft, real-time understanding of its structural health status, and effective identification of various structural damage states based on this are important ways to ensure the economic efficiency of reusable spacecraft. This paper proposes a multi-damage self-identification method for online assessment of composite material structural damage in reusable spacecraft. The method first realizes high-precision imaging of multiple damages for composite structures based on guided waves, and then clusters the damaged pixels through the probabilistic mixture model, that is, the pixel values with high damage probability in the monitoring area are matched with the Gaussian model components, so that the damage status of the structure can be automatically evaluated based on the number of Gaussian model components, their mean values and covariance. The method proposed in this paper has been experimentally validated in aerospace composite structures, and the results show that it can effectively identify the number of damages, the location of damages, and the direction of damages.