The global representation model for multi-heterogeneous social networks （MHSN） is presented， and the user space and content space association models of MHSN are established， which can be used as a reference for the follow-up research based on MHSN. Furthermore， taking the detection of localized emergencies in MHSN as an example， this paper discusses the integration method of MHSN based on content space， and conducts experimental analysis of data collection and emergency detection in Weibo and Tieba. Taking the user interest mining of MHSN as an example， this paper discusses the integration method of MHSN based on user space， and conducts experimental analysis on data collection and user interest mining of Weibo and Tieba. The results show that the proposed methods of emergency integration detection and user interest integration mining for MHSN can effectively improve the effect of emergency detection and user interest mining.