Sustainability of irrigated agriculture is based on the quality of water resources. Water Quality Indices are used to assess the suitability of irrigation water, however there are certain uncertainties when samples are to be classified near the boundary between two classes. Hence, the objective of the present study is to classify the groundwater for irrigation purpose in Tumkur district, Karnataka, India, using Fuzzy comprehensive evaluation approach for the crisp classification. The methodology of this study includes collection of 104 groundwater samples, assessment of hydrogeochemistry, and classification of groundwater by conventional and Fuzzy-logic technique. Hydrogeochemistry by Piper plot indicates mixed Na-Ca-HCO3 type and Gibbs plot indicates the influence of rock-water interactions. The water classification by conventional irrigation indices such as Electrical Conductivity, Sodium Absorption Ratio, Kelly Index, Percentage Sodium, Residual Sodium Carbonate and Magnesium Hazard showed that 2%, 0%, 86.5%, 40%, 25% (post monsoon) and 4%, 2%, 81%, 38.5%, 4% and 19.2% (pre-monsoon) of groundwater samples were not suitable, respectively. As various indices indicated dissimilar results, an integrated conventional index was evaluated by Fuzzy synthetic evaluation technique based on the Maximum Principle Membership and Fuzzy Class Ratio (FCR) and it showed 3.8% and 0.98% of samples were classified as Not suitable (Nʹ), respectively. However, FCR method was found to be effective in dealing variation in fuzzy boundary conditions and it showed 0.98%, 1.96%, 1.96%, 1.96% samples as not suitable at 5%, 10%, 15% and 20% of degree of variation near class boundaries, respectively. The study affirms that fuzziness in the data near class boundaries, consequently influencing the final classification.