This paper presents a Chinese microblog hot words extraction algorithm based on massive data Filtering. Firstly, it chooses the user behaviour characteristics and text characteristics to create user behavior models, and filters massive data to create topic-trees by a fast algorithm based on rules. Then, it uses hot words extraction algorithm to get the hot topic of topic-trees by word frequency feature. The experiment results show that the proposed algorithm can reduce the scale of the input data, with keeping lots of important information to extract hot words.