In this study an automatic cluster is done on data froma a Waster Water Treatment Plant to obtain specific knowledge of an ill-structured domain like biological wastewater treatment process. The whole process requires the supervision of the expert of the process. This clustering is done using KLASS, which allows to deal simultaneously with numerical and symbolic variables in the description of objects. From this automatic classification, we obtain a set of clusters that can be labelled as typical operational states. All the knowledge and information acquired will be constitute the initial library of a case base reasoning system that together with expert system constitute a decision support system for this WWTP.