In conventional reliability prediction methods employed for industrial robots, accurately estimating the degree of difference in reliability levels between evaluation objects and similar products poses a challenge. This paper introduces a novel approach to predicting the reliability of industrial robots based on the Interval Analytic Hierarchy Process (AHP) to address this limitation. The methodology involves an initial analysis of the relationship between the reliability of the entire machine and that of its subsystems. Subsequently, a reliability prediction model for the entire machine is formulated. During the prediction of subsystem reliability, the reliability data from analogous products are extensively leveraged. A comprehensive analysis of differences between the evaluation objects and similar products is conducted. The proposed methodology culminates in the establishment of a reliability correction factor evaluation model. This model is computed using the Interval AHP, facilitating the synthesis of deterministic information and fuzzy information to achieve a more comprehensive understanding of reliability.