Abstract: In operation of high risk construction of hydropower projects, to control and prevent the occurrence of accidents effectively, it is vital to understand the risks and find out human factors that cause accidents. The Human Factors Analysis and Classification System (HFACS) was used in this study to survey and analyze major human factors in the occurrence of accidents in hydropower construction and develop a human factor risk assessment model based on a dynamic Bayesian network (DBN). By a quantitative analysis of human factors using this model, probabilities of accident occurrences at all the time nodes were obtained and dynamic characteristics of the risk factors and their effect on accidents over time were examined. This paper also includes suggestions for reducing the possibility of accidents in high risk construction of hydropower projects.