Abstract: Traditional analyzing methods can simply calculate the value of reliability but can not analyze its accuracy or obtain the effect of every fault factor on it. So a novel analyzing method combining fault tree with information fusion was proposed to study the uncertainty of the system reliability. Variance importance is presented to improve the accuracy of reliability. Effects on the device reliability of every fault factor are calculated, and the accuracy of the most significant factor reliability is improved to meliorate the accuracy of the system reliability. Finally, the efficiency and accuracy of the method are demonstrated through an uncertainty analysis example.