The Cauca River is the drinking water source for 1.3 million inhabitants of the city of Cali, Colombia. Although the river discharge is sufficient to handle the water demand of the city all year long, significant water pollution events cause frequent disruption to the Puerto Mallarino Treatment Plant (PMTP) and the water supply service, with substantial social and economic impacts on the city. The sources of pollution include wastewater discharges upstream of the PMTP and important sediment transport from the upstream sub-catchments during heavy rainfall events. Both situations can lead to a closure of the PMTP when the presence of a pollution plume at its intake is evident. This paper presents the design and prototype of a water quality early warning system to anticipate the peaks of pollution in the river, in order to assist the operators in taking timely informed decisions about the operation of the treatment plant. As the published experiences of early warning systems for similar water pollution problems are very limited, the approach to solve the problem using hydroinformatics technologies is worth documenting for utility companies with a similar problem.