This paper investigates the factors affecting human operators’ probability of failure when performing railway maintenance tasks. The objective is to understand the interaction of the various factors and to identify driving and dependent factors. The factors are identified through a survey of the literature and ranked using a Likert scale. The reliability of measures is pretested by applying Cronbach’s alpha coefficient to responses to the questionnaire given to maintenance personnel. An interpretive structural model is presented, and factors are classified using matrice d’impacts croises-multiplication appliquéà un classement (MICMAC). The research may help maintenance management understand the interaction of factors affecting human failure probability in railway maintenance and help management devise policies and guidelines for railway maintenance related tasks.