A health state assessment method for industrial robots based on D-S evidence theory and fuzzy theory was presented. A health evaluation system for industrial robots was established,in which some health indexes, operation indexes and performance indexes were considered. Fuzzy analytic hierarchy approach was used to deal with fuzziness in uncertainty, to obtain the membership degree of the estimate measure, and it was used as the initial probability distribution of D-S evidence theory and was modified; the modified probability distribution was fused by using the D-S fusion rule, and the results of robot health state evaluation were obtained; instance analysis was given to prove the validity of the health assessment way. The results show that this method can provide reference for robot monitoring and fault diagnosis.