Micro-blog, a kind of mini-blog with limited text length and extensive publishing sources, is a typical application of informal information communication. The authors do deep research on the micro-blog, define the micro-blog, collect and construct a corpus, and analyze contents of the micro-blog and mine deep data from social network, so as to further profoundly recognize the laws of informal communication. First, define micro-blog websites and collect information from micro-blog websites which has a trait of timeline. Second, apply the improved TFIDF algorithm to construct filter corpus. Third, apply topology attributes statistics, information entries statistics, information content mining, and social network analysis to efficiently dig out all kinds of traits in micro-blog. The experiments show that micro-blog has special traits of discover small world and degree differential matching. The website interactive rates and Alexa rankings are relevant. Netizen would pay much more attention to daily topics rather than hotspot news. In conclusion, there is no formed group communication about a certain topic, but semantic chain communication between two users.