The exploration of community features is a key issue in network science and data mining. As a vital structural characteristic, community vulnerability has been paid great deal of concern. Recent works underline that many internal and external parameters to quantify community vulnerability necessarily improve conformity with topology, but are suffering from a shortage of comprehensiveness. In this paper, we propose a novel metric, namely communication and structural heterogeneity method (CSH), designed to characterize topological information by communicability and structural dissimilarity. CSH is a global path-related strategy which is based on community dissimilarity. Furthermore, intra-link number, average communicability, topological heterogeneity in communities, as well as inter-link number and structural dissimilarity between communities are employed. Thus, a more detailed evaluation of community vulnerability is suggested. The effectiveness and accuracy of CSH are verified by empirical results in real-world networks. Moreover, the propagation dynamic SIR model and simulations of random and deliberate attack are utilized to validate rationality. Meanwhile, the correlation between node importance (vulnerability) and community vulnerability is explored through experiments. The proposed method (CSH) shows its superiority when comparing it to some state-of-the-art methods.