Public participation in water conservation projects is gaining more and more attention in the information era. Public opinion, showing the focus and interests of the public, is the basis of public participation. This paper proposes a social sensing system based on social media platforms, which employs two natural language processing technologies, namely, sentiment analysis and topic modeling. The public opinion on water conservation projects is monitored from three perspectives: public opinion intensity (POI) monitoring, topic detection, and sentiment analysis. To test their effectiveness, a case study on the South-to-North Water Transfer Project (SNWTP) in China is conducted. The public opinion data were acquired from Sina Weibo, China's largest social media platform. The results indicate that: (1) POI peaks when hot project-related events occur, and POI of direct stakeholders apparently exceeds indirect stakeholders; (2) different stakeholders have different topics of concern closely associated with their interests; (3) negative events always lead to dramatic decreases in the sentiment value (SV), while positive events only slightly lift SV. The proposed system has achieved real-time monitoring of the public opinion on water conservation projects. Consequently, it can help to improve the level of public participation and provide a valuable reference in project management and policy-making.