It’s well known that the environmental and operational variations during the service life of aircraft structure could easily introduce uncertainties to the guided wave features, making it difficult to interpret the changes of guided wave signals and leading to unreliable evaluations of the damage states. This problem has become one of the key problems that hinder the engineering applications of aeronautical structural health monitoring technology. To solve this problem, a damage monitoring method based on guided wave and enhanced split merge probability model is proposed. Firstly, a stable Gaussian mixture model (GMM) is established by an enhanced split merge modeling algorithm based on the guided wave features influenced by time-varying conditions. Then, the damage evaluation is realized by observing the cumulative migration trend of GMMs, which is measured by a best matching based Kullback-Leibler (KL) divergence. To validate the proposed method, an experiment is performed on an aircraft wing spar under changing structural boundary conditions. Experimental results show that this method can achieve stable and reliable monitoring of crack propagation under time-varying conditions. In addition, compared with the damage evaluation results based on guided wave damage index, the damage evaluation reliability is improved obviously by the proposed method.