This paper takes the form of a case study of forest fires that occurred in Indonesia from 2014 to 2019 and were reported on the social media of Twitter. The study was a corpus-assisted discourse study (CADS) using data scraping or text mining on Twitter based on the keyword “kebakaran hutan” [forest fire] and discourse analysis. The actor-network theory was used to map the actors involved. This study con-cludes that five discourses show a consistently large pattern of Twitter users responding to the problem of forest fires in Indonesia. Regarding the actors, the government takes an essential role of leadership and information arbitrage on Twitter. Seeing as it is the state’s responsibility to ensure the safety of all people, the government must appear to be the main holder of control in managing disaster information traffic to avoid irresponsible information or hoaxes disseminated by parties or actors. These results indicate that the avail-ability of information obtained from every conversation of Twitter users can be used as a study or input in the formulation of evidence-based policy about forest fires. It should be given more attention as an alternative means of solving the issue of forest fires, which has become an annual problem in Indonesia