Abstract: The framework and work flow of the decision support system are proposed for fault management in water supply system through case-based reasoning as an artificial intelligent method, and the key techniques to implement the system are also discussed. After analyzing in depth the operation process of water supply fault management and extracting the case features from historical fault records, two sets of variable weights are given to the case features in case retrieval so as to implement the two core system functions: fault monitoring and fault handling. With the optimized valve shutoff scheme, simulation model of water distribution network under fault condition and two-level optimal scheduling model introduced in, the function of fault handling can be ensured even though no historical cases can be referred to, and the case database is thus substantiated further. The feasibility and validity of the system proposed are illustrated by its operating results.