The invention belongs to the technical field of road traffic operation state real-time sensing and discloses a road traffic running state real-time sensing method based on social network information,which comprises the following steps of automatically acquiring, classifying and extracting multi-angle effective traffic information in a social network platform by using a web crawler technology, screening, analyzing and predicting, establishing an end-to-end self-learning model, and performing visual labeling on a map; learning and predicting future data from past data based on deep learning, and feeding back the predicted future data to a social network for users to share traffic information in real time. The system is not limited in spatial distribution, does not need to arrange and maintain ground sensing equipment, has obvious economic advantages, can effectively capture sudden traffic incidents, traffic incidents at specific places, temporary traffic control, newly added traffic restrictions and traffic environment information, provides quick response signals for related management departments, and provides travel decision basis for residents.