A growing world population, unrelenting urbanization and increasing scarcity of water resources are the driving forces behind accelerating global demand for water. This is resulting in compromises in quality of water used in various sectors. The negative effects of use of poor quality or polluted water have been long recognized and attempts made to quantify them in terms of agro biological effects. But, the socio-economic impacts have not been given due weightage in analyses. The effect of water pollution in agriculture includes decrease in crop quantity and the consequent quality. The study is taken up in the Bhima river basin in Karnataka, India, where the river banks are densely populated and form a fertile agricultural area. Bhima River basin was purposively selected for the study because of emerging symptoms of pollution in the river. The present study attempts to assess the economic costs of water pollution on agriculture and rural livelihoods. To assess the effects of water quality, the study area was divided into two clusters namely, polluted villages cluster and non polluted villages cluster, based on the extent of effects of river water quality. A sample of five villages on the banks of river and another five villages away from the river but with similar agro economic situations was selected. Water samples were collected from two polluted and two non polluted villages and tested in laboratory for parameters like electrical conductivity, total dissolved solids, pH, total alkalinity, total hardness, chloride, sulphate, calcium, magnesium, BOD, COD, oil and grease, total suspended solids, phosphate, fluoride, turbidity, iron, arsenic, bacterial plate count, coliform count and E.coli. Tabular analysis, logistic regression and decomposition model were used for data analysis. Results of water sample tests were compared with standard norms for drinking and irrigation. Logistic regression used to determine morbidity. It revealed that the variable Vil_c which represents whether the households belongs to water polluted or non polluted village was positive and statistically significant at 5 percent level of significance. Households which belonged to polluted villages were found to be more prone to morbidity either directly or indirectly and showed higher level of morbidity. As per expectation sign of variables, Ow_land was negative, which indicated that mere ownership of land did not mean more exposure to polluted water. However, this could be due to employing labourers to work on farms. Edu_head variable was not significant but the negative sign indicates that the head of the household was expected to improve the level of awareness of the family. Pvt_toilet value which indicated the general sanitary and hygienic conditions of the households expected to reduce morbidity was negative but not significant. Signs of variables Ow_livestock, agri_lab and mig_lab were positive but not statistically significant. Decomposition model was used to decompose the change in farm productivity of sugarcane crop between water polluted villages and water non polluted villages into the impact due to polluted water used for irrigation and that due to change in use of inputs. Sugarcane crop was chosen for the study as it was a dominant in the region in terms of acreage under the crop. Average yield difference of sugarcane between polluted and non polluted villages was 3.43 tonnes/ha. worth ₹ 6,177.6. Further, the decomposition analysis showed that the yield difference due to input use was 13.10 per cent. This implied that there was sub- optimal use of inputs in sugarcane cultivation in polluted villages. However, contribution of water pollution was lower than that of input use. The water pollution depressed the productivity of sugarcane by 0.88 per cent. Monetary effects of polluted water use have captured in terms of veterinary expenses, medical expenses and employment loss/ household. Study found that average veterinary expenditure, in polluted villages was ₹ 1,710 which was more by 34.33 per cent when compared to that in non polluted villages. On an average, the household in polluted villages spent ₹ 8,197 per household per annum on medical expenses, which was 16.26 per cent more when compared to medical expenses of ₹ 7,050 spent in non polluted villages. Findings showed that these additional expenses added to the economic burden of farm family in polluted villages, which in turn affected farmers expenditure pattern and livelihood. In case of landless labourers it posed serious problem due to lower income and greater health risk due to contact with polluted water compared to other farm categories.Average number of sick days per annum in polluted villages due to which farmers were unable to work was 27 days, which was 42.75 per cent more when compared to the situation in non polluted villages and there was loss in household wage earnings due to water pollution. Average wage loss in polluted village was ₹ 7,935 per household per annum, which was 37.76 per cent more when compared to wage loss ₹ 5,760 in case of non polluted villages. River waters are polluted world wide and there effects can be studied and controlled. Statutory measures are needed to stop the effects of growing water pollution in rivers like Bhima. Study revealed that test results of selected parameters were above permissible limits in the study period which would affect livelihoods of people using the river water. Effective measures are required to regulate pollution causing industries in upstream area and set up treatment plants for industrial effluents through strict legal enforcement. Community awareness programmes are required to maintain river health. At the village level, it is necessary to identify the sources of pollution and provide remedial measures.