Abstract: A maintenance time prediction method based on virtual reality technology is presented to improve the accuracy of underlying maintenance motion time prediction. First, the maintenance time prediction process based on virtual reality is established considering maintenance simulation and macroscopical maintenance time prediction. Then, a maintenance motion classification is proposed for maintenance simulation and motion time prediction, and according to the characteristics of each classification, the procedure is set up for predicting ideal maintenance motion time based on simulation results, motion experiments and predetermined time system theories. Finally, the qualitative indicators such as accessibility, visibility, and labor intensity of the actual maintenance are transformed to quantitative values to fix the ideal maintenance motion time and compute the realistic motion time of the actual maintenance. The method was applied to the time prediction of civil aircraft maintenance task for supporting decision making of the product structure design and system layout. The feasibility of this method is verified.