Based on multidimensional personnel information, non-police incidents, online public opinion, petition, 12345 hotline, and conflict mediation data, natural language technology and big data analysis technology are used to extract and structure the multi-source data. The structured data is subjected to event relationship analysis, and an intelligent classification model is created for conflict classification to form a conflict and dispute subject library. The personnel theme information of conflict events is input into the personnel relationship model for high-frequency and high-risk personnel analysis and early warning. The event information of conflict events is input into the high-frequency and high-risk conflict model for high-frequency and high-risk conflict analysis and early warning. In the prevention and resolution of social governance conflicts, this platform provides a problem-oriented approach to strengthen the conflict resolution capabilities at the grassroots level, providing a practical solution for resolving conflicts at the initial and nascent stages.