Abstract. This paper focuses on the problem of predicating influential message (microblog) of micro-blogging services in China. Sina Weibo is one of China's most notable online social networks. Inferring influential microblog has been playing a crucial part in many applications such as branding management and online advertising. However, many previous social network analysis schemes rely mainly on link structures between users to find influencers in micro-blogging services, ignoring the important text content created by the users with the assuming that text content is the cause of links. As the text content has rich features such as word count of a single microblog, the sentiment of the microblog text and so on, we propose a novel model based on Learning to Rank, which integrates both features of the text content and the links of the Sina Weibo users to predicate the influential microblog. Experimental results show that our model outperforms many other related algorithms, including TunkRank and TwitterRank.