Classification of industrial jobs related to occupational low back disorders risks has great significance in preventing injuries and designing workplaces. In this study, industrial jobs are classified into two categories, i.e., low risk and high risk, using grey relational analysis approach. The results on a dataset are provided, together with the comparisons in terms of classification accuracy between the proposed approach and other methods that used the same dataset, including logistic regression, decision tree, neural networks, neuro-fuzzy classification, ant colony optimisation, memory-based reasoning, and ensemble model. The proposed approach outperforms other alternative methods and yields at least 10% improvement in classification accuracy compared to the best results achieved among the earlier studies. (C) 2011 Elsevier Ltd. All rights reserved.