Risk is inherent in the very concept of maintenance as minimizing the risk of failure of an equipment is why we perform maintenance. The concept of risk-based maintenance has emerged as a tool for maintenance planning and decision making to reduce the probability of failure of equipment and the consequences of failure. In this paper, we study the purposes of predictive maintenance and we give an original classification of risk-based maintenance methodologies in literature according to criteria. These criteria of classification include the input data, the structure of the risk minimization process, the decision variables, the methodology adopted, the output data, the application field and the factors that impact the quality of risk analysis. This review shows that the choice of methodologies, input data and decision variables are highly impacted by the model of the risk minimization process adopted as well as the context and the application fields.