RESUMO
Human errors in many jobs especially in nuclear, military and chemical industries may bring disaster. Supporting this proposal there are many evidence around the world such Chernobyl disaster in 1986, Three Mile Island accident in 1979 and Flixborough explosion in 1974. Therefore, human errors' identification especially in important and complex systems is necessary and thus predicting control methods are unavoidable. Recent research is a case study performed in Zagros Methanol Company in Asalouye [Southern Pars]. The study tools were observation, interview with experts and control room operators, inspecting available technical documents, and completing systematic Human Error Reductive and Predictive Approach [SHERPA] worksheets. Analysis of SHERPA worksheet indicated that,%71.25 were unacceptable errors,%26.75 undesirable,%2acceptable [with modification],%0 acceptable. Predicted risk assessment after modification was%0, unacceptable errors%0, undesirable errors%4.35, acceptable [with modification in future]%58.55, and acceptable errors%37.1. The study showed that the current implemented method can be used through various industries like to chemical and petroleum. Based on the results of the risk level assessment after modifications, once the system weaknesses were refined or controlled we should expect a decrease in the amount of human errors causing disaster