The end-effector of industrial robots is usually controlled by the blast pipe, and the signal is driven by the program to move the solenoid valve. In order to study the risk of the blast pipe in endeffector, the fault tree analysis method was introduced, and the risk identification was carried out to find out the most important factors affecting blast pipes movement. The failure risk evaluation system of the blast pipe was established, and the fuzzy theory was used to evaluate the blast pipe risk probability value, Bayesian network was used to predict the risk probability value of industrial robot failure due to the blast pipe failure, and the sensitivity analysis of each node was carried out. The results show that the risk of the blast pipe in the endeffector of industrial robots is mainly caused by corrosion, fatigue, natural force, misoperation and thirdparty damage. In order to prevent production accident caused by the fault of the endeffector blast pipe, the control and management of risk nodes should be strengthened