Falling accidents is one of the most common accidents in construction industry. In order to clarify causes of falling accidents and find effective prevention measures, 528 domestic investigation reports of falling accidents from 2012 to 2017 were collected. Based on R language and text mining technology, 32 causes of accident were identified, including 7 key causes. Apriori algorithm was used to mine association rules among the accident causes. By visualizing the results of the association rules, the causalities of the accident causes during the falling accidents were obtained, which could provide reference for the prevention of falling accidents.