For about 10 years, many methodologies have been developed to undertake a risk analysis on an industrial plant. In this paper, 62 methodologies have been identified, these are separated into three different phases (identification, evaluation and hierarchisation). In order to understand their running, it seems necessary to examine the input data, methods used, obtained output data and to rank them in several classes. First, all the input data are grouped together into seven classes (plan or diagram, process and reaction, products, probability and frequency, policy, environment, text, and historical knowledge). Then, the methods are ranked in six classes based on the combination of four usual criteria (qualitative, quantitative, deterministic and probabilistic). And finally, the output data are classified into four classes (management, list, probabilistic and Merarchisation). This classification permits the appraisal of risk analysis methodologies. With the intention of understanding the running of these methodologies, the connections between the three defined previously criteria (determinist, probabilistic and determinist and probabilistic) are brought to the fore. Then the paper deals with the application fields and the main limitations of these methodologies. So the hierarchisation phase is discussed and the type of scale used. This paper highlights the difficulties in taking into account all risks for an industrial plant and suggests that there is not only one general method to deal with the problems of industrial risks. (C) 2002 Elsevier Science Ltd. All rights reserved.