This paper presents the state of art research progress on multilingual multi-document summarization.Our method utilizes hLDA(hierarchical Latent Dirichlet Allocation)algorithm to model the documents firstly.A new feature is proposed from the hLDA modeling results,which can reflect semantic information to some extent.Then it combines this new feature with different other features to perform sentence scoring.According to the results of sentence score,it extracts candidate summary sentences from the documents to generate a summary.We have also attempted to verify the effectiveness and robustness of the new feature through experiments.After the comparison with other summarization methods,our method reveals better performance in some respects.