<p>For the foreseeable future, China’s coal mining industry will be a situation where intelligent coal mines and conventional coal mines coexist. Coal mine accident is the main cause affecting the miners’ occupational safety, and the miners’ unsafe acts is the main cause of the accident. This study is committed to using a more rigorous method to construct the causality network of the causes of miners’ unsafe acts, and reveal the internal causes affecting miners’ unsafe acts from the perspective of probability. First, use the improved Human Factor Analysis and Classification System (HFACS) qualitative establishment of the factor system that affect the miners’ unsafe acts. Second, propose the Interpretive Structural Model-Bayesian Networks (ISM-BN) model. According to the knowledge and experience of the evaluators, the causal relationship network is qualitatively constructed by Interpretive Structural Model (ISM) method. The statistical inspection method is used to conduct quantitative inspection of the qualitative causal relationship network. Third, the two causal relationship networks are applied to the Bayesian Networks (BN). The results show that the factors most sensitive to miners’ unsafe acts were Operational environmental factors. Insufficient government supervision is the source of miners’ unsafe acts in both types of coal mines.</p>