In this study, Goal-Oriented (GO) methodology, as an important technique for reliability analysis, is used in multi-state redundancy systems. However, traditional GO operation is usually not sufficiently effective for such multi-state redundancy systems because of the complex computational problem. To solve this problem, an efficient UGF-based GO operation that combines the advantages of GO methodology and the universal generation function (UGF) is proposed. The algorithm can avoid the complex calculation process caused by the algorithm-based GO operation of traditional probability formulas, and can simplify the correction of the shared signal by regarding it as the system’s conditional probability. In addition, the further correction problem in which some dependent shared signals with a common failure are caused by other shared signals is also considered. Some application examples are presented to illustrate the proposed UGF-based GO methodology, as well as its effectiveness and practicability.