Groundwater quality variations can be assessed by classical graphical representations such as the trilinear (Piper), Durov, Schoeller, Wilcox, Stiff and similar diagrams for overall quality variations. Each one of these techniques is restrictive in some sense, and it is not always possible to trace trends in the variation. Some of these do not allow a representation of more than three ions. If two or more samples have ion concentrations that are individually proportional, then their distinctive representations are not possible and they fall for instance on the trilinear and Durov representations on the same point. Statistical approaches are available but they also require restrictive assumptions such as the independence of different ion concentrations in time or space and Gaussian probability distribution. The conventional water quality index (WQI) is strongly dependent on various correlated parameters and also identification of the suitability of the parameters is critical for accurate evaluation. In this paper various simple WQI calculations are proposed on the basis of anion-cation association matrix. Different association matrices are defined based on the association by taking into consideration mean, maximum, minimum and correlation fundamentals. The association matrix definitions of WQI help to identify a set of associations between cations and anions. The application of the methodology is presented for a set of groundwater samples that are taken during 2000-2005 years.