Traditional FTA treats the failure probability as an accurate value, and only considers “normal” and “failure” states. However, as a complex electromechanical system, industrial robot system（IRS） has a great amount of cognitive uncertainties in the reliability assessment, so it is difficult to analyze the system using traditional FTA. Moreover, the mechanical structure in IRS will deteriorate caused by varieties of reasons, so it is very limited to describe the system state only by “normal” and “failure”. Aiming at the shortcomings in the traditional FTA, this paper uses T-S fuzzy FTA to analyze the reliability of IRS. Considering the maintenance cost of parts and components, the T-S fuzzy cost importance is proposed based on T-S model, and the weak links of IRS are analyzed from the perspective of maintenance economy. The results show that T-S fuzzy FTA can describe the multiple operation states of IRS, and the analysis result of T-S fuzzy cost importance is consistent with the common-sense conclusion in the industrial robot field, which provides a theoretical basis for the reliability growth of IRS.