Aiming at the problem that the existing IPTV quality monitoring platform cannot match poor quality accounts with user in- formation, and cannot accurately push the details of users with poor quality of broadcast to front-line maintenance personnel, this pa- per implements the IPTV monitor system for poor quality treatment. It uses image recognition algorithms to identify poor-quality ac- counts in batches, automatically extract related information between multiple systems, and directly push the poor-quality informa- tion to front-line personnel according to the maintenance area. The application shows that the system effectively improves the rate of IPTV poor quality in cities and towns, saves labor costs for data extraction, and realizes multi-dimensional monitoring and analysis of IPTV poor quality data.